Report of M.I.T's Ad Hoc Committee on A.I Use in Teaching, Learning, and Research Training

Eric Klopfer et al.

Audio by Paper2Audio; with some added context

Additional context

This document situates itself within the evolving pedagogical discourse surrounding the integration of Large Language Models into high-stakes academic environments. It builds upon foundational studies in educational technology that emphasize the necessity of maintaining intellectual rigor amid the rapid proliferation of automated content generation. By addressing the shift from traditional assessment models, the work contributes to a broader sociotechnical framework that examines how elite research institutions must adapt their curricula to preserve the value of expert human cognition. The committee's recommendations echo ongoing global debates about academic integrity, suggesting a departure from punitive enforcement toward a balanced paradigm of human-A.I collaboration. Ultimately, this report acts as a critical institutional roadmap for reconciling the inherent tension between technological disruption and the enduring cultivation of foundational domain expertise.

1. Introduction

This report is a call to action.

2 Definitions

Definition 1: Generative A.I: Artificial intelligence systems capable of producing novel content, such as text, images, code, or synthetic data, in response to prompts.

Definition 2: Ad Hoc Committee: A committee formed for a specific, temporary purpose, which disbands after completing its task.

Through five intense months of meetings, research, and outreach across the M.I.T community, the Ad Hoc Committee on A.I Use in Teaching, Learning, and Research Training sought to understand the role of generative A.I in the life and educational mission of the Institute and recommend how to navigate its challenges and opportunities.

In January 2026, Chancellor Melissa Nobles, Provost Anantha Chandrakasan, and Faculty Chair Roger Levy charged us specifically to:

• Assess current A.I use at M.I.T

• Identify innovations in teaching and student assessment

• Propose an A.I use policy.

However, what we learned as a group quickly convinced us that the Institute community, particularly the faculty, must tackle a set of deeper questions about the structure, meaning, and value of an M.I.T education in an era in which A.I is one of several factors complicating the Institute's mission.

Our committee consisted of undergraduate and graduate students, faculty from every school, and staff from relevant units, including the M.I.T Libraries and the Teaching and Learning Lab. Though we brought to the assignment a broad range of experience and no fixed thesis, our brief but intense explorations led us to a strong shared view.

In this report, we:

• Highlight key aspects of the current educational landscape at M.I.T

• Share eight principles we relied on and that we hope will guide the Institute in the work ahead

• Recommend immediate and long-term actions for both instructors and the administration.

As an institution deeply identified with the birth of A.I and known for its distinctively rigorous, hands-on education, designed to produce graduates unafraid of the world's hardest problems, M.I.T has a unique role to play in this moment. We believe it also has a responsibility to lead.

We hope our report can help the Institute lean into the spirit of Mind, Hand, and Heart as it continues to define and foster the highest-quality residential education – of humans, by humans, in support of human flourishing, and for the betterment of humankind.

Committee Membership

Eric Klopfer, Co-chair; Professor, Comparative Media Studies/Writing Program

Sam Madden '99, M.N.G '99, Co-chair; Distinguished College of Computing Professor, Department of Electrical Engineering and Computer Science

Faez Ahmed, American Bureau of Shipping Career Development Professor, Department of Mechanical Engineering

Jacob Andreas, Associate Professor, Department of Electrical Engineering and Computer Science

Chris Bourg, Director of Libraries, M.I.T Libraries

Cynthia Breazeal S.M '93, S.C.D '00, Professor, Media Arts and Sciences Program

Hope Dargan '19, M.N.G '23, PhD Candidate, Department of Electrical Engineering and Computer Science

Daniella DiPaola S.M '21, P.D '26, PhD '26, PhD Candidate, Media Arts and Sciences Program

Madalina Griza '26, Department of Electrical Engineering and Computer Science

Daniel Jackson S.M '88, PhD '92, Thomas J. Perkins Professor, Department of Electrical Engineering and Computer Science

Graham Jones, Professor, Anthropology Program

Ankur Moitra M '09, PhD '11, Norbert Wiener Professor, Department of Mathematics

Caitlin Mueller '07, PhD '14, S.M '14, Associate Professor, Department of Architecture

Janet Rankin PhD '89, Director, Teaching and Learning Lab

Eric So, Sloan Distinguished Professor of Management, Sloan School of Management

Marin Soljačić '96, P.D '00, Cecil and Ida Green Professor of Physics, Department of Physics

Rodrigo Verdi, Nanyang Technological University Professor of Accounting, Sloan School of Management

Thomas Wu '26, Department of Electrical Engineering and Computer Science

$ ^{1} $https://facultygovernance.mit.edu/committee/committee-ai-use-teaching-learning-and-research-training

Staff to the Committee

Jill Bassett, Chief of Staff, Office of the Chancellor

Martha Eddison, Special Assistant and Senior Communications Strategist to the President, Office of the President

Brittany Hutchinson, Executive Assistant, Office of the Chancellor

1.1. The Landscape

This report's recommendations reflect the following insights about the educational landscape at M.I.T.

Generative A.I is everywhere already, spurring an assortment of views:

- M.I.T students use A.I frequently and pervasively – with strongly mixed feelings, from curiosity, creative inspiration, and gratitude to resignation, concern, and anxiety.

• Instructors' attitudes range from enthusiastic exploration and growing reliance on A.I to skepticism, suspicion, and “A.I refusal”—and there's a widespread desire to share experiences, ideas and techniques.

While A.I is allowing instructors to develop exciting new learning experiences and enabling students to learn and experiment in innovative ways, there have been a number of concerning effects on the life of the campus, including signs that A.I is:

- Upending foundational elements of the M.I.T educational experience, especially for undergraduates, from the p-set, the take-home exam and urop's to office hours and the study group

• Increasing isolation

• Undermining student mastery and confidence

• Eroding the “social contract” between instructors and students

• Making it much harder to assess student progress

• Challenging decades of distinctive M.I.T community norms and values about rigor, the creative friction required for learning, collaborative problem solving, and personal integrity.

Definition

Tectonic disruptions: Profound and far-reaching changes that fundamentally alter the structure or nature of something, analogous to geological plate movements.

A.I is generating both immediate rapid changes and long-term tectonic disruptions – and M.I.T needs to respond:

- Students are confused and concerned about a lack of clarity, consistency and justification about the use of A.I, within a given subject and across the curriculum

- Every subject taught at M.I.T will likely need to be reexamined and potentially revamped to make sure that how students are being taught, what they're learning, and how they're assessed are "A.I-aware."

- A.I is changing what students need to know and know how to do.

- Beyond its implications for specific subjects and disciplines, the advent of A.I demands a broader, holistic reassessment of the nature, scope and purpose of higher education today.

Definition

Disparate access: Unequal or varying availability and usability of resources or technologies among different groups or individuals.

Finally, A.I presents substantial practical concerns for our community, from data privacy and confidentiality to questions around disparate access, bias, fairness, and accountability. A.I's explosive growth and ascendance also raise important questions for society, from the environmental impact of A.I data centers, to the use of intellectual property and training data, to the overall human impact of the technology and the industry – and M.I.T needs to engage with those questions too.

Other educational institutions are grappling with similar questions around A.I and generating interesting ideas – but no one seems to have it all figured out.

Three notes on the words we use:

N.B. In this report, “Instructors” includes faculty and everyone else engaged in teaching at M.I.T.

Most references to “students” apply to both undergraduate and graduate students, except in a few obvious places, as when we refer to the General Institute Requirements (G.I.R's) or to participants in the Undergraduate Research Opportunities Program (urop).

• “We,” “us,” and “our” sometimes refer to the members of the committee, and sometimes the whole of M.I.T. The difference should be clear from the context.

2. Guiding Principles

We start by defining eight principles we relied on and that we hope will guide the Institute in the work ahead.

2.1. Be humble

Definition

Mutual adaptation: A reciprocal process where two or more entities adjust and change in response to each other's evolving states or behaviors.

Some technological innovations emerge gradually: As society and technology evolve in concert, mutual adaptation softens the impact. The computer – A.I's precursor and key enabler – fits this pattern. Other innovations land more abruptly, becoming socially consequential before individuals and institutions have time to adapt.

Society tends to peg the “birth” of a new technology as the point when it becomes readily usable. By that measure, generative artificial intelligence was “born” with the release of ChatGPT in late 2022. Public engagement with generative A.I is therefore less than four years old. In that time, it has amassed more than a billion users, and the companies selling A.I technology have come to dominate the headlines, the stock market, and public consciousness.

In other words, A.I is progressing across almost every domain and on a timescale too compressed for society to properly observe and analyze its impacts and then gradually adapt.

This suggests our first guiding principle: We offer our proposals in a spirit of humility. Course corrections – perhaps even major ones – will be inevitable as the technology continues its relentless evolution and the Institute experiments and learns.

2.2. Be bold

Yet uncertainty can't be an excuse for inaction. This is not a moment for patches and duct tape. The challenges A.I presents in teaching and learning call for a bold strategic response – everywhere, and especially at M.I.T. With our Social and Ethical Responsibilities of Computing program completing its seventh year, we are uniquely positioned to find ways to employ this new technology for the benefit of society, and for our students in particular.

A.I also presents extraordinary opportunities, from unprecedented possibilities for individualized tutoring and coaching to a dramatic acceleration and revamping of research in many disciplines. Seizing these opportunities deserves and demands boldness too.

Bold thinking is especially important because our students will go on to help shape the intellectual, ethical, and technical direction of our society – and soon. We owe them a deep engagement in rich and constructive uses of A.I, and a sophisticated understanding of its potential and its drawbacks. Their M.I.T experience should prepare them with the wisdom to help determine how and where A.I is used for the betterment of society and the world at large.

$ ^{2} $https://computing.mit.edu/cross-cutting/social-and-ethical-responsibilities-of-computing/

2.3. Put humanity front and center

Facing a technology that already has such immense capabilities – built on, modeled after, and in many ways now exceeding human powers – the Institute's fundamental challenge and most important goal must be to nurture and protect our shared humanity, and to value the M.I.T community, its members, and their flourishing above all.

In a listening session with instructors, we learned that some were considering using A.I agents as research assistants instead of hiring undergraduates as urop's. One can see the case for speed and efficient use of resources (especially now, when research resources are so constrained). But if those criteria come to dominate our decisions, we all have to ask, "What is it that we are here together to do?"

As a community, we need to keep in mind that although research is central to M.I.T's mission, it's more than an end in itself; on our campus, research is also an apprenticeship, a means of training the next generation of researchers to continue our work and drive our fields of inquiry forward. Using research as an opportunity for learning-by-doing may produce seeming "inefficiencies," but that's a feature, not a bug.

A focus on our humanity should also alert us to the fact that common current remedies for the problems A.I creates can risk damaging the relationships between and among teachers and students. For instance, many instructors recounted that having to “police” unauthorized A.I use was harming their connection to students (a dynamic made worse by the fact that, as we learned, A.I detection software is quite unreliable). For their part, students fear being wrongly accused of A.I cheating and are increasingly frustrated by instructors' use of A.I in areas that demand a human touch, such as grading, assignment creating, and feedback. Such an underground river of mutual suspicion is no foundation for a healthy classroom.

2.4. Lean into learning

Many problems and assignments used in M.I.T classes to reinforce learning and assess student progress can already be accomplished by A.I. Unrestricted A.I use by students can make some traditional assessments less reliable as indicators of individual learning, potentially weakening confidence in grades and credentials. For educators, this radical shift in norms and expectations can feel profoundly disorienting.

But as a community, what should worry us most is that many uses of A.I deprive students of the opportunity to learn.

The threat A.I poses to familiar ways of teaching and testing may be a blessing in disguise – because the changes are too sudden and severe to ignore. As the faculty on our committee can attest, for at least two decades, educators have lamented the strain on teaching and learning from fragmented attention, ubiquitous devices, a narrow preoccupation with grades rather than learning, and a weakening sense that students and teachers are joined in a common intellectual project.

As an institution that prides itself on the power of its distinctive educational recipe, it is up to M.I.T to seize the opportunity of this moment: to make sure that the undeniable changes imposed by A.I become a tipping point, forcing us to deal decisively with the forces eroding our shared educational mission.

Leaning into learning means creating a new “social contract” between teachers and students. All of us who teach at M.I.T will need to be prepared to help students understand both that the process of education is necessarily a productive struggle, and that the most important product of their education is not a G.P.A or a diploma but themselves: their personal growth and intellectual maturity and the development of their own imagination, insight, and judgment.

Definition

Metacognitive abilities: The capacity to think about one's own thinking processes, including planning, monitoring, and evaluating one's learning strategies.

Instilling these attitudes needs to become a central task for every educator, so that our students know not only what they should learn, but also how they should learn and why learning matters. We need to help them develop metacognitive abilities to think about thinking, to engage in reflective practices, and to enhance their sense of personal agency. This will require both dedication and fresh preparation on the part of instructors.

2.5. Teach with intentionality

Definition

Pedagogical assumptions: Underlying beliefs and principles that guide the methods and practices of teaching and learning.

Most classes taught at M.I.T will require careful review and, in many cases, substantial adaptation to survive an encounter with A.I without serious disruption. A.I will undermine the pedagogical assumptions that guided the choice of content, the methods of instruction, and the ways of assessing student performance.

In short, it's time to reflect deeply and extensively on what we teach and how we teach it. Future class structures, assessments, and policies must be built with high intention and a clear sense of purpose, not merely tweaked in reaction to the immediate realities of A.I. This does not mean, however, that every course must change in the same way.

Definition

Backward design: An instructional planning framework that begins with defining desired learning outcomes and then designing assessments and learning experiences to achieve those outcomes.

The time-tested technique known as “backward design” offers a useful framework here. For instance, rather than beginning by asking whether A.I should be allowed or prohibited in a specific subject, educators would begin by defining the purpose of the learning experience itself: what students should come to know, be able to do, and learn to value. Assignments and assessments can then be designed to measure and encourage those outcomes, using A.I if it's helpful but not if it isn't.

The instructor time and effort required to redesign, reassess, or intentionally preserve existing structures is likely to be significant. But that process of pedagogical reflection and reasoning is valuable in itself. Intentional teaching promotes intentional learning; when instructors make clear why A.I is permitted, limited, or required, students are more likely to understand the learning that's being protected or developed.

In this way, backward design can help M.I.T preserve the purpose of its education while adapting honestly to the realities of A.I. (Encouragingly, M.I.T's Teaching and Learning Lab is well-equipped to help instructors master this tool.)

$ ^{3} $https://tll.mit.edu/teaching-resources/course-design/backward-design/

2.6. No one size fits all

Definition

A.I fluency: The ability to effectively and ethically use, understand, and interact with artificial intelligence tools and systems.

No single approach to A.I in education will serve M.I.T well. A poetry seminar, a mathematical proof course, and an architectural design lab logically entail different relationships to A.I. In some, using A.I to generate creative work or solve a problem defeats the purpose of the exercise; in others, A.I fluency is already part of professional practice, and avoiding it would leave students shortchanged.

Similarly, a first-year student building foundational skills and judgment stands in a different relationship to A.I than doctoral candidates accelerating literature review in a field they deeply know. In short, a uniform Institute-wide rule would inevitably be too permissive for some contexts and too restrictive for others.

This is not to say that every instructor should have to invent A.I policy from scratch. M.I.T should provide a shared framework – including a common policy menu, disclosure expectations, and accountability standards – within which departments and instructors can make choices. Departments are a natural unit for this work, since students experience a major as a coherent progression and would benefit from consistency about when A.I is welcomed, when it is restricted, and why.

2.7. Augmentation not automation

As A.I-enabled technologies become more capable, it will be possible – and tempting – to offload more “thinking” tasks to them; not surprisingly, students told us the temptation was greatest when they feared they would miss a deadline. There are early signals, however, that overreliance on chatbots can have a range of significant negative consequences – diminishing critical thinking, weakening memory, eroding confidence, and undermining mastery.

2 Definitions

Definition 1: Cognitive surrender: A psychological state where an individual ceases to engage in effortful cognitive processing, opting instead to rely on external aids like A.I when faced with challenges.

Definition 2: Pro-worker A.I: An approach to developing A.I systems that aims to augment human capabilities and enhance worker effectiveness rather than replace human labor.

Getting the right answer from a chatbot can create the illusion of learning – but it can also trigger “cognitive surrender”, where students fall back on A.I at the first hint of struggle. In a recent paper, three noted M.I.T economists – Daron Acemoglu, David Autor and Simon Johnson – argue for what they call “pro-worker A.I.” They make the case that instead of designing A.I systems to replace workers, businesses should develop A.I that allows human beings “to be more effective at their existing tasks, tackle new tasks, and acquire new expertise.”

In the same spirit, collectively, we need to shape M.I.T's use of A.I in education to be "pro-learner" – not to replace the "hard fun" of thinking, learning, creating, and problem solving, but to expand what's possible for students to think about, learn, and solve. In short, A.I should be used to augment and enhance curiosity, creativity, and learning, not automate them.

Definition

Deskilling: The reduction or elimination of the need for specialized skills in a workforce, often due to automation or technological advancements.

Importantly, this issue is social as well as psychological. Just as automation threatens job loss, deskilling, and the increasing isolation of workers, so too the reliance on chatbots in education displaces peer-to-peer connections like study groups and decreases interactions between students and instructors. If we want augmentation to carry the day instead, we need to see it as more than an abstract, “hoped-for” ideal; it's a concrete practice that our pedagogies can help promote.

This leads to our final principle.

$ ^{4} $https://osf.io/preprints/psyarxiv/yk25n_v1

$ ^{5} $https://www.brookings.edu/articles/building-pro-worker-ai/

2.8. Think beyond the classroom and the campus

A.I should be integrated into M.I.T education in ways that enhance, rather than diminish, students' capacities for growth as whole human beings. In The Culture of Education, Jerome Bruner, a leading psychologist of education and a pioneering figure in cognitive science, argued that learning is not simply a technical matter of transmitting content from teacher to student, but a process through which students learn how to interpret the world, find their place in society, and develop a sense of themselves as capable persons.

If learning is (mis)understood primarily as content delivery, then highly responsive A.I systems may seem capable of replacing teachers, mentors, peers, collaborators, and the interpersonal dimensions of learning altogether. But while this attitude may be increasingly common, it's shortsighted. Education is a cultural practice through which students learn to make meaning, exercise judgment, form identities, and participate responsibly in community.

Definition

Tacit expectations: Unspoken or implicit understandings, norms, or assumptions within a community or social setting.

Much of what our students gain from M.I.T is never spelled out in a syllabus or an assignment; it's what they learn from living and working on our campus in each other's company – the tacit expectations, habits, relationships, and values that inform how they learn to solve problems, exercise judgment, persevere through difficulty, and become members of an intellectual community. "Residential education" is powerful in part because it happens everywhere: in residence halls, living groups, sports teams, arts groups, clubs, and so on.

A kind of social contract

During our discussions, we came to think of the alignment among these elements as a kind of social contract: a shared understanding among students, instructors, and the Institute about the purposes of education, and the rights we enjoy and responsibilities we owe as members of an intellectual community. Effective institutional responses to issues under consideration must therefore attend not only to policies and rules governing A.I use, but also to the norms and culture that give those policies meaning.

Definition

Transactional model: A view of education where learning is seen primarily as an exchange of inputs (assignments, tuition) for outputs (grades, degrees), focusing on efficiency over deep understanding.

A fundamental danger, as we've discussed, is that A.I can allow students to bypass learning. Equally concerning is that students may internalize a transactional model in which assignments are outputs, teachers are evaluators, peers are optional, and knowledge (or an M.I.T degree) is an optimizable commodity to be acquired or produced as efficiently as possible.

Such a mental model will not remain confined to the classroom. It will shape how students come to understand work, collaboration, and social responsibility, and they will carry that mindset with them out into the world. However, sociable or responsive A.I systems become, they cannot substitute for the relationships and practices necessary to grow and mature as a human being.

M.I.T should therefore approach A.I not simply as a tool for improving educational efficiency, but as a force that must be governed to support the broader purposes of education. Students should learn to use A.I productively and fluently, but also in ways that deepen their capacity to participate well in their classrooms, laboratories, workplaces, families, and communities.

The goal is not to shield students from A.I, nor to preserve older educational forms for their own sake. It is to ensure that A.I use supports the development of people who can think critically, act with initiative, work productively with others, and understand the consequences of their choices in a world shared with 8.3 billion other human beings.

$ ^{6} $https://www.hup.harvard.edu/books/9780674179530

3. Recommendations

This section details the changes we see as necessary for M.I.T to prepare our students for a new world. Based on the Institute's longstanding values, our recommendations are not a checklist of individual initiatives that can be implemented one at a time, bit by bit, but rather a set of substantive changes that must be undertaken in concert.

We recognize that serious change takes time. Given the impacts already affecting our community, however, the required changes should be implemented on two timescales: those that happen immediately and those that begin immediately, but require further study and planning.

Artificial intelligence – in the form of L.L.M's and other generative A.I technologies – presents M.I.T with profound challenges and intriguing opportunities.

Already these technologies can produce credible solutions and provide reasonable responses to almost any written assignment in our undergraduate curriculum, including essays, math and science problems, proofs, and coding assignments, and their power will only grow.

Concerning impacts

Because many students are choosing or feel pressure to shift to learning and problem-solving with A.I, in less than three years these technologies have driven major shifts in campus culture, including decreased attendance at office hours, reduced participation in online discussions, and, as we heard anecdotally, a drop in in-person study groups in dorms, libraries, and other study spaces. These issues have presented themselves suddenly and dramatically, creating a clear sense of urgency. They also land at a time when higher education is facing other challenges, and M.I.T itself is considering broad curricular changes emerging from the findings and recommendations of the Taskforce on the Undergraduate Academic Program (T.F.U.A.P).

Intriguing opportunities

While these challenges are pressing, A.I also offers exciting opportunities for learning and for creating. Many instructors told us that A.I helps them develop customized, interactive learning tools that allow students to explore subject content with more depth and for instructors to create learning experiences for their students that are new or newly tailored to each student. With a little guidance, even instructors who are not proficient in software development can customize A.I agents to support a subject or research project.

Students can create large-scale software projects with the limited timeframe of classes that would not have been remotely possible before. They can also use A.I to analyze data, conduct research, and build tools to amplify their expertise and pursue projects that benefit them and society at large. The potential of these technologies to augment work across campus is immense.

The recommendations we offer below aim to help the M.I.T community navigate A.I's challenges and seize its opportunities, while reflecting the principles laid out in Section 2. For instance, Humility means that we must create administrative processes for continuous evaluation and revision rather than assuming any change we make today will be adequate in the future. Boldness requires us not simply to try to patch the existing system to limit or counteract the impacts of A.I on our students and systems but instead to redesign learning experiences, assessments, and curricula in ways that help redefine the future of education. Putting humanity front and center means that we will seek uses of A.I that strengthen rather than weaken the value of an M.I.T residential education. Leaning into learning means teaching with intentionality, revisiting what we need students to learn, and aligning assessments with those desired outcomes, to preserve the productive struggle. essential to a learning-by-doing education. No one-size-fits-all means that we should develop tailored frameworks for adapting to A.I rather than uniform A.I rules for all students, instructors, or departments. Augmentation over automation means that we prepare students to use A.I fluently and in ways that preserve agency, judgment, integrity, and human connection. Thinking beyond the classroom and the campus requires that we anticipate the skills that our students will need for success in work and life, as community members, future leaders, and creators of the next generation of A.I technology.

$ ^{7} $https://gue.mit.edu/tfuap/

Three broad areas of work

We believe the circumstances call for a substantial reshaping of the student experience. While the life of an M.I.T student will no doubt look different in 10 or 15 years, as a community we have a responsibility to make sure that it remains immediately recognizable as an M.I.T experience. Paradoxically, preserving what's most important will require deliberate change.

We group our recommendations into three sections, briefly summarized below. Broadly speaking, we recommend that, through the combined efforts of the faculty and the administration, M.I.T:

Adapt Educational Processes for an A.I-Aware World (Section 3.1)

- Campus-wide and as soon as possible, review what our students need to learn. Led by instructors and departments, this “A.I-aware” process will inform the design of classroom activities that are more social and experiential, and in tune with innovative assignments and assessments that provide feedback on those key student outcomes. This will also require an investigation into the nature of grades and grading.

• Facilitate this process by providing instructors with dedicated resources (for example, additional T.A's) and by adapting physical spaces for experimentation, community-building, and A.I-free interaction and assessment.

- Create “communities of practice” at M.I.T that allow members to share challenges and findings, as well as tools and techniques to improve teaching and learning.

Center people, community, and the residential experience (Section 3.2)

- Reinvigorate and reclaim the residential experience based on shared experiences, transparency, and a focus on in-person activities.

- Thoughtfully integrate A.I literacy and responsible and ethical A.I use as a foundation and throughline for student work, attending closely to how it connects to specific disciplines. This will demand technical proficiency, of course, but it hinges on adopting a human-centered approach to A.I as an augmentative tool.

Build processes, teams and tools to enable continuous reflection, iteration and improvement (Section 3.3)

• Establish roles, processes, and ongoing committees to address these challenges

- Collect feedback and data, stay apprised of research and best practices, connect with peers and learn from them

- Provide technical and educational support to enable all of this.

In addition, we identify a set of concerns about the wider impact of A.I on society, jobs, the environment, and more, that also bear on our on-campus A.I policy.

3.1. Adapt Educational Processes for an A.I-Aware World

Generative A.I affects what our students need to know and be able to do (our learning outcomes), how our students achieve those outcomes, and how we evaluate and assess what they have learned. Because A.I's ability to competently complete M.I.T-level assignments makes it difficult to assess student progress based on out-of-class work, instructors urgently feel the need for new assessment strategies.

While we make several recommendations about assessment, we urge instructors to do more than simply try to “A.I-proof” their classes. In particular, the ways that A.I changes the social aspects of learning – such as how students interact with T.A's and engage in office hours and study groups – present an even deeper challenge.

3.1.1. Revisit course goals

Definition

A.I-aware assessments: Evaluation methods designed to account for the existence and potential use of A.I tools by students, aiming to measure genuine learning.

Before considering how to construct A.I-aware assessments, instructors should reconsider their goals for student learning in every subject they teach. What should students know or be able to do by the end of the course?

The learning goals for a given subject may or may not be affected by the availability of A.I, but they should be “A.I-aware”; that is, they should recognize that A.I exists in the world, that instructors may direct or permit it to be used in or out of class in structured ways – and that students may also seek to use it without permission.

We heard that the greatest concern for many instructors is: When A.I makes it possible to offload the cognitive work of learning, how can we assess what students actually know and understand?

Definition

A.I-resilient: Designed or adapted to maintain their validity and effectiveness in the presence of A.I tools that could otherwise compromise traditional methods.

In the era of A.I, some traditional learning goals may merit rethinking; for example, do the majority of our students need to be able to write complex programs by hand? Individual instructors and departments will need to assess whether longstanding goals are still important and, if so, develop A.I-resilient ways to evaluate them.

While learning goals that involve lab, project, or performance-centered work may not require much adjustment, generative A.I may offer novel ways to help students reflect on, practice, or prepare to perform. For instance, an instructor reported to us that, for students learning to serve as mediators, providing personalized, course-specific A.I coaches eliminated the awkwardness of practicing public speaking in front of others, which substantially increased students' willingness to practice, which in turn increased their skills.

A.I may also enable instructors to devise learning goals that were previously impossible, such as understanding or working in new ways with very complex texts, engineering artifacts, or large software systems.

$ ^{8} $https://tll.mit.edu/teaching-resources/course-design/backward-design/

3.1.2. Ensure durable learning through new course policies, structures, and forms of assessment

Definition

Cognitive friction: The mental effort and challenge involved in deep thinking, learning, and problem-solving, considered beneficial for robust understanding.

M.I.T's curriculum is famously demanding. On top of that, many students pack their schedules with high unit loads and multiple extra curricular activities. These competing demands on their time drive students to prioritize efficiency – and nothing could be more efficient than automating work through A.I. But if students give in to that tempting option, they cheat themselves of the cognitive friction and productive struggle necessary for actual learning.

A.I has both created and revealed a mismatch between established learning objectives and familiar forms of assessment. At M.I.T, we have long used problem sets, long-form written assignments, take-home exams, and projects completed outside of class to allow students to solidify their knowledge through practice and demonstrate what they've learned. A.I saps these tools of their value for both teaching and assessment.

Instructors already sense it: In the face of A.I, they have to change how they evaluate students. Already, many are increasing the weight of exams in their grading or asking students to write or code during class time. However, these tactical solutions come with a cost: For instance, overemphasizing in-class evaluations means reducing students' incentive to invest themselves in the difficult, time-intensive p-sets and projects it takes to build the ladder to mastery. By definition, shifting assessments to time-limited class periods reduces how much thought and deliberation students can put in. If we want students to care about and know how to create and recognize worthy work – work of scope, rigor, creativity and thoughtfulness – quick, high-stakes evaluations embody the opposite of the signal we want to convey to them right now.

Instead, instructors need to revisit what they really want students to know and devise assessments that foster, or even include, the kind of productive struggle that builds durable understanding and mastery.

We urge instructors to consider forms of assessment that are less vulnerable to A.I, and more valuable for learning, such as oral exams, semester portfolios, and out-of-class assignments paired with in-class conversations. This likely means that resources such as T.A's and class time will become more central to evaluation and also raises the question of whether we should find ways to limit class sizes.

As described in Section 3.3 below, we also hope that the Institute will create efficient ways for instructors to share new assessment ideas and best practices that may emerge.

3.1.3. Emphasize experiential and project-based learning

To match the shift towards assessment methods that aren't vulnerable to A.I, instructors need to increase the role of experiential and project-based learning. To encourage this kind of creative teaching, M.I.T should support the development of teaching skills and practices for all instructors, and recognize contributions in this domain.

Definition

Experiential learning: A pedagogical approach that emphasizes learning through direct experience, reflection, and application, often involving hands-on activities and real-world projects.

Ironically, in some cases A.I itself can help open the door to experiential opportunities. For instance, since A.I excels at tasks like coding and some types of design, instructors can now assign projects that are much more ambitious: in M.I.T's capstone level software engineering class, projects that students could reasonably undertake in a semester often had to be limited in scope and unrealistic in key respects. Today, it's entirely reasonable to expect students to use A.I coding tools to build near production-quality software artifacts in a single academic term. This in turn can make room for deeper experiential learning: students can explore how different designs work in the real world, reflect on cases where certain designs outperform others, and explore how the systems they build work in realistic settings.

Opportunities extend well beyond computer science: Architecture students are using A.I to experiment with new ways to visualize and rapidly test their ideas, beyond what's possible with traditional representational skills. This expands the range of creative projects they can undertake (though it only increases the importance of building fundamental conceptual and technical skills, judgment and social reasoning, to ensure students remain in charge of their ideas.)

Definition

Mens et Manus: Latin for 'Mind and Hand,' M.I.T's motto, signifying the integration of theoretical knowledge with practical application.

In a recent Wall Street Journal interview, Harvard economist Rafella Sadun observed that workers who will succeed in the A.I era are those “who can creatively think about ways in which A.I can generate value, quickly translate ideas into action, and who are ready to revise their intuition through experimentation and validation. Tinkerers and doers thrive in this environment” – an obvious match with M.I.T's hands-on ee-thoss and focus on societal impact. As A.I enables our students to tackle more sophisticated and challenging tasks, one way to live up to our Mens et Manus motto will be to expand the scope and real-world application of the projects and endeavors we assign.

Finding new opportunities for collaborative, in-class projects provides significant opportunities for growth beyond mastering the particular content. Properly structured and supported, such projects can build students' strength in collaboration, communication, problem-solving, and emotional intelligence. These durable human skills remain among those that employers prize the most. We would do well to provide more opportunities for our students to build them.

A.I can also enhance these experiences, offering students new ways to analyze data, conduct research, and build tools to amplify their expertise and pursue projects that benefit them and society at large. The potential of this technology to augment work across campus is immense.

3.1.4. Build structured in-person social learning into subjects

Because A.I tools are disrupting longstanding patterns of social learning at M.I.T, every subject should include a regular in-person social component (not just sitting in lecture or recitation and quietly taking notes). Instructors should intentionally structure such interactions to achieve desired learning objectives and maintain quality, even in large classes.

Examples of structured social learning include:

- Group projects with weekly staff check-ins and deliverables that assess both individual and collaborative contributions

• Group problem-solving sessions guided by a lab assistant or T.A

• Feedback discussions structured around a class rubric

- In-class discussions facilitated by a trained T.A or professor, where individual participation is graded.

10 get the most value from in-person interactions like these, instructors should clearly explain on the first day of class why they are important for individual learning and building community. Courses that include extensive group projects should offer practical guidance and structures for how to successfully work together.

Although the presence of A.I is disrupting social learning on campus, A.I tools can also offer real benefits for student learning, including timely help, individualized feedback, accessibility, and support for students who might otherwise be stuck, as well as just-in-time ways to help groups collaborate.

Unfortunately, these benefits become counterproductive if they allow students to bypass the human settings where they would learn how to work with others, communicate their ideas, receive criticism constructively, build confidence, develop judgment, and act as members of a community.

M.I.T should therefore treat A.I not only as a pedagogical or technical challenge, but as a critical opportunity to renew residential education around human presence, shared work, and meaningful mentorship.

$ ^{9} $Chen, T.-P., & Lahart, J. (2026, June 9). Economists weigh in on the future of work and AI. The Wall Street Journal. https://www.wsj.com/tech/ai/economists-weigh-in-on-the-future-of-work-and-ai-f59311e9

3.1.5. Preserve and expand out-of-class research and career experiences

With its launch in 1969, M.I.T's Undergraduate Research Opportunities Program (urop) broke new ground by creating a systematic way to give undergraduates hands-on experience with frontline faculty-led research. Today, urop is a signature part of the M.I.T experience, directly engaging 93% of undergraduates and 58% of faculty.

urop and other out-of-class research and learning experiences offer students obvious practical benefits such as academic credit and paid work. But the broader benefits – including personal development, connections with mentors and peers, and career exploration – are even more important. urop makes an eloquent case for the purpose and value of residential college education.

M.I.T should preserve, renew and ideally expand its commitment to urop and other learning experiences that emphasize mentorship, collaboration, and learning by doing; find ways to make them more accessible; and consider extending them beyond the lab.

We also encourage the Institute to explore related efforts to enhance the value of the M.I.T residential experience, such as establishing co-op programs (which the Institute is actively exploring ^{11} ), allowing urop-like experiences to satisfy degree requirements, or expanding industrial/co-op based graduate research programs. This should include working with organizations on campus that already support experiences like these, including the P.K.G Center for Social Impact and misti.

urop's, research assistantships (R.A's), and other forms of student participation can give students entry into new intellectual communities: joining a lab, learning from graduate students and postdocs, collaborating with peers at different career stages, presenting work in group meetings, contributing to papers, traveling to conferences, and gradually coming to see themselves as members of a field. Students learn not only methods and technical competencies, but also how research questions are formed, how judgment develops, how mistakes are constructively interpreted, how credit is shared, how disagreement is handled, and how knowledge is collectively produced.

One note about urop in particular: While undergraduates can provide faculty with useful research labor, that is not the point of the program. It exists to educate. This is why the possibility that faculty may find it advantageous to replace novice researchers with A.I agents warrants special concern.

As A.I systems become a cheaper or more efficient replacement for urop's or R.A's, students could lose access to the relationships, practices, and shared forms of work through which belonging, confidence, judgment, and professional identity are formed. The danger is not only that students would have fewer opportunities to conduct research, but fewer opportunities to become participants in research communities.

$ ^{10} $https://urop.mit.edu

$ ^{11} $https://orgchart.mit.edu/letters/announcing-mit-co-op-planning-committee

3.1.6. Reconsider grades and incentives

We do not advocate limiting the grades instructors assign through grade rationing (e.g., imposing a limit on the number of A's that can be given). The committee felt that this would be counterproductive to campus culture and to the Institute's efforts to navigate the challenges of A.I. Due to the intense societal focus on credentials and grades, students who are set on maximizing their G.P.A have a strong incentive to use whatever means they feel are most effective to achieve that goal. Rationing top grades would intensify the temptation to cut corners on actual learning by increasing reliance on A.I.

As instructors explore new forms of assessment and experiential learning activities, M.I.T should take this opportunity to consider what role grades play in our overall system, and if the current approach could be improved. For instance, if instructors mainly respond by putting more weight on in-class exams, there's a risk of narrowing what the M.I.T credential has long signaled: that M.I.T students are capable of difficult, independent, and thought-intensive problem solving, not just acing exams on paper.

The committee encourages M.I.T to explore alternative systems of grading and assessment, including those that work in other countries (such as the U.K system, which uses percentages to express relative mastery against an expert standard), and newer paradigms such as competency-based and mastery-based assessments . As one thought experiment, the committee discussed the idea that if M.I.T did not have grades, many of the incentives around A.I cheating would disappear. Note that M.I.T already refrains from the common practice of awarding summa/magna/cum-laude diplomas, because we believe that earning an M.I.T degree stands as distinction enough.

Grades are far from the only way for students to demonstrate mastery or gauge their own progress; real-time feedback during an oral exam or presentation, or written comments reflecting on a project submission, likely give students more useful information about their mastery of concepts than they could glean from any grade.

With an increased focus on projects and experiential learning, instructors should consider whether grades are the only, or even the best, mechanism for signaling mastery in all contexts. Alternatives could include elevating the standing of portfolios of projects as a showcase of student achievement in every field. Already, many employers are less focused on applicants' grades than on their performance on internal assessments, such as the ability to answer difficult interview questions or demonstrate problem solving and mastery through custom exercises.

3.1.7. Expand in-person spaces for labs and in-person evaluation

In several ways, our recommendations call for more face-to-face activities, from hands-on learning and collaborative lab projects to in-person assessments. All of them require physical space.

Several instructors expressed a desire for collaborative spaces where students can complete extended assignments or exams. These might be fully analog spaces or spaces with A.I-free (or A.I-limited) computers. Many departments have lab spaces that could support collaborative activities but lack resources to staff them full time. (In fields that use expensive and possibly hazardous equipment and materials, self-service labs are often not possible.)

Collaborative hands-on spaces are useful for teaching, learning, and community building across the disciplines, not just in the traditional lab sciences. We urge M.I.T to invest in creating and staffing such spaces.

$ ^{12} $https://en.wikipedia.org/wiki/Competency-based_learning

3.1.8. Provide A.I Use Policies, with Justification

Students report that the A.I guidance they receive from instructors is often confusing and unclear. What's more, the rules and the way they're conveyed vary widely from instructor to instructor.

While we recommend against imposing a one-size-fits-all policy on A.I use, students are anxious for clarity about A.I use: in any given course, they want to feel sure about when, where, and why A.I is prohibited, allowed, or required.

Therefore, instructors, and perhaps departments, should make sure that every M.I.T subject has a clear policy about the use of generative A.I, posted prominently in the syllabus and on the course website.

To make this workable, as soon as possible M.I.T needs to develop a clear and consistent menu of guidelines for instructors and departments to choose from and adapt as necessary. Departments will likely want to take a coordinated approach so that guidelines and rationales are well understood and largely consistent across a given major.

We encourage instructors to use a standardized format to communicate this information, so students can easily understand what is and is not permitted, course by course. Appendix B below provides an example policy.

A.I policies should include a rationale. While many courses already have an A.I use policy, we recommend that all instructors make sure that their policies include a clear rationale, tied explicitly to the learning goals of the given course, for why A.I must, may, or may not be used.

For example, if an instructor wants to ban the use of generative A.I tools instead of simply declaring that A.I use is a form of cheating, it's more effective to explain how generative A.I tools shortcut students' ability to learn the fundamentals of the course or to practice solving the kind of problems they will encounter in exams, future classes, and the real world. This approach is more likely to foster self-awareness and a healthy culture around A.I use.

Similarly, instructors who want to encourage A.I use in a particular assignment should make it clear why using A.I is important to the educational experience. For assignments that feature A.I, instructors may wish to include exercises in which students reflect on when and how A.I helps or harms their learning, thinking, and morale, and how it changes or expands the work they produce.

Finally, when allowing or encouraging A.I use, instructors should make sure students understand that L.L.M's can fabricate facts and citations, code generation tools can produce incorrect or insecure code, and image generation tools can produce biased, offensive, or otherwise inappropriate outputs.

A.I is only a tool, and students are responsible for all work they submit, including any inaccurate, biased, offensive, or unethical content produced with the assistance of generative A.I.

3.1.9. Exercise Caution with A.I Detectors and Online Exam Platforms

Many tools purport to detect A.I use, and some are reasonably accurate when given a piece of purely A.I-generated text. However, most real-life situations are more nuanced: If a student uses A.I to complete an outline or edit a section, these tools are unlikely to uncover it.

Definition

A.I detectors: Software tools designed to identify text or content that has been generated by artificial intelligence.

No doubt the tools will improve. Nevertheless, we recommend against relying on A.I detectors, as it risks an arms race in which students respond to automated detection by using increasingly powerful “A.I humanizers” to remove signals that A.I detectors are cued to catch. The result: a lot of effort on both sides that in the end serves no one.

A.I detection systems may also mistake the writing of non-native English speakers or neurodivergent students for text generated by A.I. Even low rates of false positives can put students on edge and cause serious individual consequences. More generally, stepping up “policing” around A.I use builds an adversarial atmosphere of distrust between instructors and students, which understandably hurts students' motivation and morale.

Some instructors have asked if M.I.T will provide so-called "lockdown" browsers for conducting evaluations of students. These are online testing environments that take over the computer during an exam, preventing access to any online resources outside of those provided as a part of the test. The committee recommends that M.I.T study such tools, but notes that the current generation is buggy, error-prone and feels like surveillance. At least for now, in-person proctored exams – the current norm at M.I.T – are a better choice in most cases, though they will require appropriate in-person spaces.

Alternatives to these heavy-handed technological solutions are more likely to build trust and support student learning. Options include working on writing in class by hand for early ideation/drafts, creating regular project deadlines or meetings to show project progression, preventing massive amounts of work from piling up near the final deadline, and allowing instructors to give feedback at various stages (not just on the end product).

Other technical tools can also be helpful. For example, instructors can require students to do their work on platforms that capture a history of versions, and to submit the history along with their work. This can provide useful process evidence – for example if a student were to submit an assignment within a few minutes though comparable work would typically take hours. And students themselves often find these tools useful for reflecting on how their ideas have evolved.

Preparing our disciplinary system

While we hope students will follow A.I policies for their own benefit, M.I.T needs to be better prepared to handle serious violations. M.I.T's existing academic integrity framework, and the Committee on Discipline (C.O.D) processes that enforce it, were built on a model of authorship in which the work students submitted was either their own or someone else's. Generative A.I blurs the lines. A student who uses an A.I tool to brainstorm an approach, debug a function, tighten a paragraph, or generate a full draft is not "copying" in any traditional sense, and faculty in our listening sessions consistently reported uncertainty about where the line now lies, both in formulating their own course policies and in deciding which cases to refer to the C.O.D.

M.I.T needs to make sure that Institute policies clarify what evidence would be required to bring an academic integrity case forward when A.I is involved; the C.O.D itself does not consider A.I detector output alone sufficient.

Students should not feel policed. Durable change will require instructors to be as clear as possible about their expectations and students to understand A.I misuse as an unacceptable deviation from shared peer norms and community values rather than a violation of an arbitrary bureaucratic rule.

3.1.10. Support responsible experimentation in the curriculum

There are many unknowns about how best to integrate A.I into the curriculum and how to design experiences to avoid its pitfalls. We encourage instructors and departments to pursue coordinated curricular experimentation in both these areas. It might also make sense to offer variants, for example by developing both rigorous "A.I Light" or "A.I Heavy" classes or pathways through the curriculum.

Some students and instructors prefer not to use A.I (for a variety of reasons). We encourage instructors to take such student reluctance seriously and, where the subject matter allows it, to suggest a pathway that keeps the use of A.I to a minimum.

Creating learning experiences that meet clear but perhaps evolving learning objectives, incorporating A.I where applicable, and always acknowledging its presence – all while maintaining high standards – won't be accomplished by superficial tinkering. Instead, it will require bold, imaginative systemwide experimentation and evaluation. Our work on this committee demonstrates the shared will among students, instructors, and staff from every corner of the Institute to creatively address this epochal challenge together.

The committee heard from many instructors that the administrative process for making even small changes to curriculum requirements is time-consuming and burdensome, making it hard for M.I.T to respond to, let alone anticipate, fast-moving, far-reaching disruptions.

We urge faculty governance to review its processes this fall to ensure the presence of efficient pathways for curricular experiments that enable more rapid exploration, paired, of course, with thorough evaluation of the results. It will also be critical to raise awareness of new and existing pathways across departments, sections, and programs, so they can immediately begin exploring A.I-aware substitutions and alterations in their curricula as soon as possible – or we will be left behind. At the same time, of course, it's essential to maintain the kind of long-term coherence across our curriculum that has been a hallmark of M.I.T education (see Section 3.3 below).

3.2. Center people, community, and the residential experience

A.I is disrupting the social fabric of our campus – and the wider world. The backlash against A.I in commencement speeches in the summer of 2026 was just one expression of the discontent many feel about A.I itself and the role Big Tech now plays in our society.

An M.I.T education has always meant being challenged by a deep intellectual community that champions creativity and rigor. We must preserve the transformative power of that experience for our students.

A.I has arrived at a moment when the social foundations of campus life are already under strain. Across higher education, students report high levels of anxiety, depression, loneliness, and disconnection. Confidence in institutions has weakened, political polarization has made campuses more fragile, and the post-pandemic recovery of in-person communities remains incomplete.

A.I is therefore not arriving in a socially neutral environment. It is imposing new pressures on a community still working to restore the foundational habits and attitudes that residential education depends on: showing up, building trust, enthusiastically joining shared efforts, and living with an easy sense of mutual obligation. A.I could accelerate this erosion. Or it could be the impetus for a deliberate rebuilding.

The challenge for M.I.T is therefore neither to preserve existing practices unchanged, nor to reject A.I tools that may support learning and research. It is to act decisively to ensure that A.I augments the human communities through which education happens, rather than quietly replacing them – so that the machines serve us and not the other way round.

M.I.T should use this moment to ask what forms of presence, collaboration, mentorship, and shared intellectual life are essential to residential education – and then design A.I policies, pedagogies, and research practices that strengthen them rather than allow their erosion, replacement, or collapse.

$ ^{13} $https://healthymindsnetwork.org/wp-content/uploads/2024/09/HMS_national_report_090924.pdf

3.2.1. Define and communicate the value of residential education

As a residential institution, M.I.T needs to clearly define the value of bringing students to campus and engaging avidly in face-to-face discussions, dialogue, and problem solving. No one reading this report was educated without these basic human interactions. But we cannot take for granted that incoming students will appreciate their value.

This is the core of M.I.T: In the spirit of “Mind and Hand,” we work together on real projects that require deep thought and careful, rigorous work.

It's clear that unthinking, “opportunistic” use of A.I tools to complete assignments diminishes what students learn and damages the learning communities they belong to. Learning works when it's both challenging and social; knowledge is built through cognitive friction, whether that's disentangling the steps of a mathematical proof with your study group, adjusting an experiment over and over until it works, or having a spirited argument with a peer (rather than getting “the” answer from A.I). That's why it matters for students to go to college!

Students need to believe that doing the work themselves is in their self-interest. It will take more than top-down A.I use policies to inspire that conviction. As a community, we need to cultivate a campus culture in which A.I restrictions that instructors impose clearly support student learning, personal growth, and professional success; and students have a shared culture that values and fosters residential learning and discourages excessive reliance on A.I.

M.I.T's official mission and values unite us all. But our daily decisions, conscious and unconscious, are largely shaped by the unspoken norms and values of our immediate communities – classes, study groups, departments, labs, offices, teams, clubs, living groups and more.

For our students, it's these in-person communities that foster the social relationships – formal and informal – that allow them to grow, thrive and prepare for their lives beyond M.I.T.

Employers prize M.I.T graduates for excellent problem-solving and technical skills. But they also want people with the skills of human connection. By prompting students to turn to A.I instead of to their fellow students, T.A's, or professors, these technologies offer the illusion that human interaction is inconsequential. But learning with and from other people simultaneously supports learning and builds the human skills that make the world go 'round.

$ ^{14} $https://web.mit.edu/about/mission-statement/

$ ^{15} $https://web.mit.edu/values/

3.2.2. Strengthen social connection and personal wellbeing

A.I threatens the shared social fabric of M.I.T and creates new risks to the well-being of our students. M.I.T needs to invest in creating shared norms, establishing new community rituals, and addressing new academic, social, and emotional challenges introduced (or exacerbated) by A.I.

While our core recommendations center on policies for the classroom and A.I training, we encourage the Institute to take many other steps to build the kind of community that A.I can't replace.

In the past, the technical General Institute Requirements (G.I.R's) provided not only a shared academic foundation and problem-solving skill set for first-year undergraduates but also an intense bonding experience around a common core. But this core is fracturing as students rush to take major-specific classes in their first year and put off some of the G.I.R's until senior spring. Students' other shared rituals – campus preview weekend, orientation, convocation, the brass rat, and commencement – are widely spaced milestones. Outside of their research groups or labs, graduate students have even fewer shared rhythms and rituals.

Anecdotally, it appears that the rapid adoption of A.I may already be affecting student well-being and creating new long-term challenges. Some students are becoming dependent on A.I for schoolwork or emotional support; losing their sense of agency, purpose, and confidence; and worrying about their career prospects in an A.I-saturated world. Instructors and other academic staff should be prepared to recognize A.I-related academic, social, and emotional challenges and to direct students to appropriate support.

M.I.T should continue to invest in and expand its first-year learning communities programs. But we also need to engage many more people across campus in conversation about shared norms and why community matters. One example could be a series of panels that bring together students, people from industry, instructors, and staff, to talk about where they come from and what they love: not just what they do, but how they have handled obstacles and made hard choices in life and career. Other ideas include campus-wide celebrations of human skills and accomplishments, “Tech Free Times” (periods during which M.I.T would not schedule classes, office hours, or meetings and instead support activities centered on in-person connection), and broadening M.I.T Reads, perhaps at the department or residence level, to encourage more people to participate in collective reading, reflection and discussion.

3.2.3. Encourage Instructor Disclosure Around Their Own A.I Use

Just as M.I.T needs norms around students' A.I use, the community also needs to agree and adhere to norms on the appropriate use of A.I by instructors. If instructors plan to present students with content substantially generated by A.I, or to use A.I for some aspect of evaluation, grading, or feedback, we strongly recommend that they be transparent with their students about how and why A.I is being used.

While A.I can be a remarkably useful tool to help instructors improve lecture notes, generate slides, and create animations or simulations, students made clear that its over-use in the classroom can make them feel that teaching is not a priority for their instructors. A common refrain: “Why should I bother coming to class or doing the work if the teacher is just going to give an A.I-generated lecture?” In a poll, most students told the Tech they are uncomfortable with the idea of A.I teaching assistants.

Our listening sessions reinforced how alert students are to instructors' behavior: Students notice when instructors are clearly using A.I for slides, feedback, or grading while imposing restrictions on student use, and they perceive it as a double standard. This points not only to a risk but also to an opportunity: by modeling responsible use of A.I, instructors can help reinforce the new social contract our community will need to adapt to an A.I-infused world.

With respect to evaluation, while A.I can be a good way to check assignments against a rubric or automate some aspects of evaluation, many students also have a justifiably strongly negative reaction if, after investing their own time in an assignment, the only feedback they receive is from a machine. An alternative way to use A.I graders is to provide them directly to the students along with the assignment, as a way for students to get useful feedback on their performance rather than a way to determine final grades. If we view assignments as a critical tool for learning, both the instructor and the student need to invest their own thoughts and efforts in the process.

$ ^{16} $https://gue.mit.edu/tfuap/

$ ^{17} $https://firstyear.mit.edu/first-year-advising/learning-communities/

$ ^{18} $https://libraries.mit.edu/mit-reads/

3.2.4. Teach Effective, Responsible, and Ethical Use of A.I

In the fall 2025 Tech Survey, more than two-thirds of students who responded felt that A.I would be important in their careers, yet only 25% felt that M.I.T was adequately preparing them to use A.I.

The effective, responsible, and ethical uses of A.I are distinct – but interrelated. Students need ways to build strength in each area. We recommend immediately integrating these topics into undergraduate and graduate orientation, and infusing them throughout the curriculum as soon as possible, so community members can build a shared sense of competence in assessing and evaluating A.I systems.

- To use A.I effectively, students need to learn how to specify a problem or prompt, how to verify an output, when a model is likely to hallucinate, and when not to reach for A.I at all.

- Using it responsibly begins with grasping the difference between augmentation and automation when it comes to education: when A.I helps a human do better work, versus when it replaces the human (and thereby eliminates the learning that would come from the human doing the work). Responsible use also entails knowing how to disclose A.I's contribution honestly across coursework, theses, internal reports, and published research.

- Ethical use requires each of us and all of us together to explore issues our community has not yet worked through, including judgments about the provenance of training data; the potential for bias; the risk of homogenization in outputs; environmental and resource cost; intellectual property issues; and the line between assistance and misrepresentation of authorship.

These themes can be instilled through the curriculum in a range of ways: Capstone projects in a given major could include A.I-related components, such as opportunities to reflect on the limits of A.I and its appropriate use within the discipline. Departments may also choose to develop “A.I-I-M” (A.I Intensive in the Major) courses, and similar foundational courses for grad students, to teach students how the use of A.I meshes with practices in their discipline. Communication Intensive subjects should also provide instruction and practice around how to use and when to avoid A.I in writing, including the risks of hallucinated citations and of the homogenization of language and ideas. Graduate advisors and research groups should incorporate discussions of appropriate use of A.I in conducting and publishing research.

$ ^{19} $https://thetech.com/2025/11/25/llm-survey-results

3.2.5. Recognize and Mitigate Negative Impacts of A.I

As we learned from both listening sessions and survey results, many people at M.I.T object to the way most A.I models were trained – by reading vast amounts of data from the internet, often without the authors' explicit consent. Some community members decline to use A.I at all because they object to how the data has been harvested without benefit or attribution to its creators.

M.I.T needs to recognize the broad range of attitudes and perspectives around A.I, from enthusiasm to skepticism to outright rejection, and create educational experiences that will allow all students to thrive academically and personally. M.I.T should also encourage research and policy efforts to mitigate the negative impacts of A.I.

While corporations aggressively push their A.I products in the education sector, including on our own campus, people are becoming more concerned with the growing size and power of leading technology companies. This is particularly true for students, who have grown up under the sway of powerful social media firms. They would like to see both greater responsibility and greater transparency around the societal, environmental, and social impact of A.I tools.

3.2.6. Acknowledge A.I Use in Theses and Other Research Work

Student research is a hallmark of an M.I.T education. Many of our students contribute to research publications, and student theses, made publicly available through the M.I.T Libraries' M.I.T Open Scholarship repository, share our research with the world. Because these publications not only disseminate knowledge but also launch careers, they should reflect the value and integrity of an M.I.T education.

All theses should include a statement about how A.I was used in the production of the thesis. A.I should never be listed as a co-author. Most journals and conferences also require acknowledgment of A.I use, and some prohibit certain use cases, so researchers should check guidelines before preparing manuscripts for publication. The M.I.T Libraries provide guidance on how and when to cite A.I tools.

Just as in the classroom, norms play an important role in how A.I is used in research, particularly in the preparation of manuscripts and theses. What uses of A.I in research are acceptable? How do we educate researchers in the practice of discerning when A.I output is useful and factual?

Standards and practices in different communities will vary. But we believe that all M.I.T researchers should view A.I as a tool under the direction of a human who is responsible for verifying the accuracy of its output.

3.3. Build processes, teams and tools for continuous reflection, iteration, and improvement

The previous sections outlined many changes, both immediate and longer term, that M.I.T will need to make to adapt to A.I. Given the startling rate of change in A.I, however, the solutions we offer today must continuously evolve. Our recommendations in this section are intended to address this need for deliberate ongoing adaptation.

$ ^{20} $https://dspace.mit.edu/communities/6fc02cc2-0d14-4023-8a6f-d9900d0c4302/search

$ ^{21} $https://libguides.mit.edu/cite-AI-tools

3.3.1. Establish an Ongoing A.I and Education Committee

It was clear to all of us on the committee that, although our report is finished, M.I.T's work on this subject has only just begun.

To put our recommendations into practice, M.I.T should establish an ongoing committee. It will serve to monitor the on-and off-campus impacts of A.I; evaluate changes in A.I models and technology; and incorporate evolving evidence from the economy, learning sciences, and campus experience into policies and practices.

The committee will be the primary group responsible for strategy, monitoring, evaluation, and policy coordination, working closely with the A.I Leads, the A.I Fellows, and the A.I Implementation Team (whose roles we describe in the sections below).

3.3.2. Create School/College-or Department-Level A.I Leads

We recommend appointing A.I Leads in each school and the college (or possibly each department) to help with local A.I curricular planning, adaptation, and adoption.

Their responsibilities could include shaping tailored A.I policies, planning how to adapt each unit's course offerings in light of A.I, and developing new unit-level education and assessment models such as competency-based learning, portfolios, or both.

3.3.3. Fund A.I Fellows and an A.I Implementation Team

We also recommend that M.I.T fund an A.I Implementation Team and A.I Fellows – staff members, postdocs, urop's, and graduate students with the skills to help instructors adapt their courses to A.I in ways that make the most of the latest A.I technology and are backed by the principles of learning science.

Including experts in technology, A.I, and learning sciences, the Team and the Fellows would work at three levels: supporting individual instructors; developing campus-wide tools; and serving as implementation experts with an "ear to the ground" to study how these tools are being used across campus and to what effect.

Transitioning from longstanding classroom practices and assessments to a new set of A.I-aware or A.I-enabled practices and assessments presents a daunting task for any individual instructor. Doing this at scale across the Institute will be a massive undertaking.

If we expect instructors to make more than incremental improvements, we need to offer them guidance, direct support, and community. Given that existing instructor and curriculum support units like the Teaching and Learning Lab are already at capacity, the success of this endeavor will require expanding the pool of expert staff.

3.3.4. Create an A.I Pilot Fund

Curricular changes are not easy, and many of our instructors, particularly in the largest classes that most immediately need to adapt to A.I, already feel overworked and unsure of how to proceed. To support A.I-aware revision of course objectives and methods across campus, M.I.T needs to provide substantial resources.

We recommend that M.I.T and departments/schools/the college create an A.I Pilot Fund that instructors can apply to for resources (such as A.I credits, T.A's, urop's, and summer support) to do A.I-enabled projects, explore the impact of A.I on pedagogy, and create deliberately A.I-free experiences.

3.3.5. Provide ongoing training and instructor support

In addition to material resources, instructors need ongoing opportunities to learn how A.I is changing teaching, learning, and their own work. Many instructors described having relatively limited experience with A.I and feeling unsure about how students are using it and how to use it themselves. Others told us that using A.I has helped them develop interactive learning tools, create animations for slide decks, polish lecture notes, and reduce various aspects of course management busywork, while also creating new learning experiences for their students. The tools have already progressed to the point that even instructors with no background in developing software can produce impressive results.

To help instructors navigate both the risks and the opportunities, we recommend that M.I.T follow the lead of M.I.T Sloan and develop a set of resources for educators, including:

• Holding regular "lunch-and-learn" style seminars where instructors can hear from colleagues about how they are using and adapting to A.I. Instructors across the Institute were eager for such "communities of practice" both within and across disciplines.

- Creating online and in-person training on the use of A.I for teaching and administrative tasks, such as how to develop animations and simulations, use coding tools, build task-tracking systems, and so on. While these might incorporate components on ethics and A.I safety, we urge M.I.T to avoid “checking the box” by adopting generic third-party A.I training programs and instead invest in ways for our own community to provide these resources.

- Institute-sponsored annual or bi-annual workshops on A.I and Education for both the M.I.T and wider academic community to convene and discuss developments and best practices.

Some of these changes can be facilitated by the A.I Implementation Team and Fellows mentioned above. The Teaching and Learning Lab, which focuses solely on supporting residential education, can play a key role in helping to organize and conduct some of these activities.

3.3.6. Develop metrics

To understand how A.I is reshaping our community and classrooms, M.I.T should begin tracking metrics around A.I use, campus engagement, student satisfaction, and post-graduation feedback. This could include additional A.I-focused questions in the biennial M.I.T Quality of Life survey, or perhaps other metrics, such as additional questions in course evaluations about student/staff engagement and A.I use.

$ ^{22} $https://mitsloanedtech.mit.edu/ai/

3.3.7. Ensure equitable technology access

M.I1 should continue to provide a "model agnostic" A.I platform such as Parley, that does not tie our community to one model provider, and ensure that this platform provides the level of access that our community needs.

Different generative A.I models vary significantly in their performance and behavior on different tasks. Commercial A.I providers like OpenAI, Google, and Anthropic all offer for-pay hosted commercial services, with the highest tier of plans costing as much as $200/month (as of June 2026). Some researchers reportedly spend thousands more per month on pay-per-invocation requests ("A.P.I Access") for coding tools and other use cases. These commercial providers continually renew their models to offer state-of-the-art performance.

For the M.I.T community, the result is an uncomfortable inequity: Many students here can spend $200 or more per month out-of-pocket to access cutting-edge A.I systems; many of their peers cannot. Some students can thus access tools that can complete substantial course projects or solve very difficult math and coding exercises overnight; others must make do with less advanced models, even if course policy allows unrestricted A.I use. This imbalance has the potential to create enormous disparities in course performance.

Similar concerns apply outside the classroom, to administrators and researchers whose productivity is being measured against colleagues with access to more or less powerful models.

As an important first step in addressing access and privacy concerns, I.S&T offers the Parley System, which gives everyone at M.I.T access to a range of different models, both commercial and open, and provides each user with up to 30/month of free credits to use them. A.P.I access for use with coding tools was released in the Summer of 2026.

The access Parley provides is very helpful, but it may not be sufficient for all M.I.T classes, particularly in software engineering where industry is rapidly adopting these tools. Some users have also expressed concern that the $30/month "budget" for each user is too low. We recommend that M.I.T continue to survey the community and consider providing expanded access for certain types of uses (e.g., specific A.I-focused classes) where students need additional resources. Additionally, we have heard from several groups that even now, they still pay for commercial models because subscription plans effectively subsidize token usage relative to Parley's per-token pricing.

Definition

Agentic systems: A.I systems designed to operate autonomously, make decisions, and take actions to achieve specific goals.

Because agentic systems and tools are increasingly part of A.I practice, whatever systems we employ should provide access to agentic coding and working tools, in addition to chat interfaces. We also need to ensure that we provide adequate access to computing resources that students require for classes and projects where they are training or fine-tuning models, considering options of both on-campus and cloud-hosted resources.

3.3.8. Protect sensitive data and preserve model choice

We recommend that M.I.T continue to invest in access to state-of-the-art tools that allow students, instructors, and staff to get work done, without having privacy concerns about the chats, data, or applications they're using. M.I.T must take responsibility for ensuring this layer of privacy while remaining model agnostic.

One concern about commercial A.I tools is that sensitive data may be incorporated into their training sets. Most plans, especially paid ones, allow users to deliberately disable the capture of their data. for training purposes – but the default is often to record user sessions for training. This leads to the possibility of private, sensitive data from M.I.T users finding its way into the answers that these models provide.

M.I.T's I.S&T provides a good reference around allowed use for different types of data with these commercial A.I providers.

Some students and faculty have concerns about how their publications (including student theses, journal articles, and books) might be used to train A.I. The M.I.T Working Group on Scholarly Content and Generative A.I has produced guidelines and resources on this topic.

We strongly support Parley's model-agnostic approach, and we urge M.I.T not to commit to any single commercial A.I ecosystem, while maintaining first-class access for all. We would also like to see M.I.T offer more hosted access to local A.I models where users can issue requests without fear of training data capture by commercial providers. Recent open-source models that can be run on local GPUs provide capabilities similar to commercial models on many tasks and will continue to improve, but to run locally, these largest open models require very large GPU clusters.

$ ^{23} $https://parley.mit.edu

3.3.9. Establish privacy, logging, and auditing policies

A.I systems are not just tools for working. Since the very earliest days of the technology, people have used them for emotional and psychological support. This even occurred with one of the earliest systems, the M.I.T-created eliza chatbot (though ironically it was mainly designed to demonstrate the shallowness of human-A.I interactions).

Today, members of the M.I.T community share information with A.I systems that includes the most personal and sensitive aspects of their identity and experience, posing questions about their health, relationships, emotional lives, goals, aspirations, and anxieties. Our students are part of the generation most likely to use A.I for this kind of personal advice.

Because Parley is an M.I.T-administered system, one opportunity it provides is to audit usage of A.I models; in theory, I.S&T can see every request from users to Parley and responses from the backend A.P.I models.

M.I.T therefore needs to establish a clear policy about what types of logging and auditing will be allowed. There are competing interests: On one hand, users may want to use these models to ask private or personally sensitive questions with the expectation that their data will be confidential. On the other, instructors may find it valuable to see how their students are using A.I chat sessions or logs to see if learning goals are being met. M.I.T may need to explore anonymization technology, or establish a practice so that classroom assignments where logs are shared with instructors are clearly labeled and advertised.

A few questions M.I.T should resolve: How should we retain and store data? How should we monitor (or intervene in) A.I interactions when users raise the possibility of harming themselves or others, or otherwise show signs of psychological distress? How can we best communicate these policies (and their implications) to the campus community? These issues need to be discussed openly and resolved transparently.

$ ^{24} $https://ist.mit.edu/ai-tools

$ ^{25} $https://provost.mit.edu/wp-content/uploads/2026/08/Scholarly-Content-Gen-AI-WG-at-MIT_-Recommendations.pdf

$ ^{26} $https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/

3.3.10. Monitor A.I Costs and Environmental Impact

Many community members shared serious concerns about the environmental costs of A.I and the impact of data centers on communities where they're built. Issues include data centers' energy and water consumption, the impact of their construction, and the materials used to build their hardware. These have been echoed in popular literature and were the subject of a report at M.I.T. Both training and inference incur energy costs, and the rapid response of models, as well as the constant churn of training newer and more powerful models, has an environmental impact.

M.I.T should address this both through campus practice and through research.

While we believe it is essential that M.I.T provide access to A.I tools, we also recommend that the Institute invest in ways to provide information about the actual environmental and financial costs of using A.I. This could include publicizing estimates of the environmental cost of different models on Parley, or encouraging the use of A.I tools with lower impact. We also encourage M.I.T, to the extent possible, to audit its own A.I footprint.

There is ongoing work on campus to measure the environmental costs of data centers and find ways to reduce them. The impacts are often difficult to quantify, both because companies keep this data private and because they are diffuse, encompassing everything from construction to training costs spread over the life of the inferences. Many groups at M.I.T – including the M.I.T Energy Initiative, the M.I.T Climate Project, and M.I.T Climate and Sustainability Consortium – are exploring innovative answers around the future of data centers and communities. M.I.T could also contribute in important ways to designing A.I models that use much less energy.

4. Conclusion

We hope this report reflects the complexity, urgency, and nuance of the challenges and opportunities that A.I presents for M.I.T and helps the Institute community come together to take action.

A.I has the potential to help our students be more productive, learn more efficiently, and build skills that can help them make a better world. What's more – though this exceeds the scope of this report – A.I's extraordinary capabilities are already transforming the practice of scientific research, on our campus and elsewhere. But A.I also imposes many disruptive risks to the process of learning, the practice of teaching, and the social fabric of residential education.

A strong shared view

A striking finding from our work as a committee was that – although our members came from across the Institute, including M.I.T's five schools and the college, and from different professional perspectives and generations – we all shared the same fundamental concerns about the risks A.I inflicts on our students and our community, and the same urgent determination that M.I.T lead in pioneering a new model of education that celebrates human wisdom and connection.

M.I.T's Unique Position

Because of M.I.T's longstanding leadership in A.I, its role as one of the world's preeminent institutions for research and education, and its long tradition of hands-on education, the Institute is uniquely positioned to lead higher education in this high-speed, high-stakes adaptation to the A.I era.

This is not an optional exercise. We must demonstrate how to integrate A.I thoughtfully and deliberately into the classroom, the research enterprise, and the experience of residential education in ways that ensure learning, advance discovery, and prepare students for the future.

There is a lot to do.

Starting now, the M.I.T community must embark on a set of changes to make our educational approaches more A.I-aware. That need not mean incorporating A.I into everything we do, or even most things. But none of us can afford to be oblivious to the fact that A.I is influencing what our students learn and how they learn it.

$ ^{27} $https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117

$ ^{28} $https://mit-genai.pubpub.org/pub/8ulgrckc/release/2

For students

In this work, we will serve our students by emphasizing what makes the residential experience so transformative and doing more of it; by rethinking what they need to know and revamping our assessments to match; by teaching them to use A.I appropriately within their disciplines; and by helping them develop the skills, judgment, and attitudes to succeed beyond M.I.T in an A.I-saturated world. We must foster a community that values and makes the most of the shared in-person M.I.T experience and that challenges students to join in the productive struggle of learning.

For instructors

Definition

Communities of practice: Groups of people who share a concern or a passion for something they do and learn how to do it better as they interact and share their knowledge regularly.

The Institute also needs to make sure that instructors know how to use A.I and when it's best not to. Though instructors themselves must take the lead in weighing what students need to know now and devising "A.I-aware" assessments, M.I.T – and perhaps especially the departments – likely need to provide considerable support, coordination, and guidance. We must build "communities of practice" within and between schools, so instructors can learn from each other about how to use, adapt to, and defend against A.I. This in turn requires the Institute and the departments to allow and support experimentation with, about, and around A.I in the classroom, and to provide ways to capture and share those experiences.

Over time

Some changes will take time. We need to find ways to provide students with appropriate A.I learning experiences from the time they arrive at M.I.T through graduation. We need to continually monitor and study the technical and durable human skills that industry demands, and ensure that we are preparing our students appropriately.

We must deepen our scientific understanding of how A.I affects learning and make sure our educational practices reflect that evolving knowledge. And while we reap the benefits of A.I, perhaps especially in the realm of research, we must actively work to reduce its harms.

A roadmap to the future

With humility, we offer this report as a roadmap to help the Institute navigate this transition immediately and over time. The work will require thoughtful commitment and focused effort from the entire M.I.T community – students, instructors, staff and leadership. The goal: to preserve and enhance the distinctive transformative power of an M.I.T education.

It is not too much to say that M.I.T's mission depends on it. Fortunately, our work as a committee convinced us that the people of M.I.T are intensely interested in meeting this challenge and uniquely equipped to do it, so that the Institute can continue to develop in its students "the ability and passion to work wisely, creatively, and effectively for the betterment of humankind."

A. Committee Process

Our committee consisted of undergraduate and graduate students, faculty from every school, and staff from relevant units, including the M.I.T Libraries and the Teaching and Learning Lab. Given the rapid pace of change in this area, the committee's work was brief but intense—we met weekly during the spring of '26 and conducted surveys, analysis of peer policies and practices, a number of listening sessions with the M.I.T community (including separate sessions for graduate and undergraduate students), as well as meetings with a range of groups working on A.I on campus, including I.S&T, the Committee on Discipline and the Office of Student Conduct and Community Standards, M.I.T Open Learning, the school councils, and a number of individuals we felt had particular insight and expertise in A.I and education.

A.I Use: None of the text of this report was generated with A.I. We did supply early drafts of the report to ChatGPT to identify redundant / overlapping sections. Some of the graphs and statistical summaries in Appendix C were generated with Codex (OpenAI's A.I coding tool).

Members of the committee also used A.I to synthesize and present data and reports (e.g., summaries of other schools' published positions on A.I) to inform committee discussions.

B. Sample A.I Policy for Syllabi

We recommend that the campus agree on a menu of standard policies for A.I use in classes, and that every class have a clearly stated policy selected from this menu (and adapted as necessary).

As described in Section 3.1.8 above, once instructors choose an option from the menu, they need to convey it clearly to students through the syllabus and perhaps on the course website, along with an explanation of the rationale behind the choice: Given the goals for what students will learn in the course, how does the selected A.I policy support and encourage that learning?

The idea is to make sure students are clear about what is and is not allowed in each course or assignment and, more broadly, to help them understand what the costs and benefits are to their own growth.

Below is a proposed 4 option menu which may be applied to a whole course or specified for individual assignments. Using the traffic light icons can help make this system easy for students to decipher.

B.1 Unrestricted GenAI Use

Policy: Students may use any GenAI system for any purpose on assignments in this course.

Best suited for:

Courses where the primary assessment is done via in-class (A.I-free) evaluations or assignments where GenAI use does not undermine the learning objectives, such as ambitious software-system projects that are beyond the capabilities of current tools or where evaluating and integrating A.I output is itself part of the work.

Instructors may wish to ask students to include short per-assignment disclosures describing how they used GenAI, possibly including transcripts or session records when appropriate. Instructors may also wish to ask students to reflect on how A.I use did or did not help with the assignment or their understanding.

B.2 Limited GenAI Use: Support Tool Only

Policy: Students may use GenAI to support learning, brainstorming, editing, debugging, or generating explanations, but may not use it to produce full or substantial assignment solutions. This is similar to the way that a student might ask a classmate or T.A for assistance. Assignments in this category should clearly specify which stages, processes or components allow and disallow use.

Prohibited uses may include:

Copying full problem prompts into GenAI tools, asking GenAI to solve assigned problems, or submitting A.I-generated text, code, equations, or analysis as one's own.

Best suited for:

Courses where independent problem solving is central, but where A.I can reasonably serve as a tutor, editor, or study aid. Students must clearly distinguish their own work from A.I-assisted work and cite A.I-generated material or ideas. This can be good for courses that already have clear collaboration norms and want a simple, intuitive A.I policy aligned with existing academic-integrity expectations.

Note that in our experience students and faculty may differ as to what “support” means — for example, some people feel that using A.I to write an outline or A.I to generate diagrams is more than “support.” For instructors with strong opinions about this, policies should clearly state what types of uses are and are not allowed.

We expect that this is the appropriate policy for many of our courses, but also requires the most thought in specifying boundaries.

B.3 Required GenAI Use

Policy: Students are required to use GenAI to complete the assignment in the manner specified.

Students must use the designated tool or class of tools, follow the prescribed workflow, and submit any required documentation of A.I interactions.

This may accompany one of the other categories, as there may be components that require A.I and others where it is disallowed.

Best suited for:

Assignments where learning objectives include A.I-assisted programming, prompt engineering, critique of model outputs, human-A.I collaboration, or evaluation of A.I systems.

B.4 GenAI Use Strictly Prohibited

Policy: GenAI may not be used in any form for the course.

Because policing out of class use as a reference aid or support tool is essentially impossible, this policy is difficult to enforce reliably and may create risks of both undetected violations and false accusations.

C. Survey Results

In this section, we summarize the results of three recent surveys of the M.I.T community around A.I: a survey our committee designed in spring 2026, a set of questions we added to the M.I.T Quality of Life Survey (also conducted in spring 2026), and a survey conducted by the M.I.T Tech (mostly focused on undergraduate and graduate students) in fall 2025.

C.1 Spring 2026 A.I Usage and Attitudes Survey

We conducted a survey of the M.I.T faculty, instructional staff and students regarding their A.I usage and attitudes. We received 1632 responses, representing a 12% response rate. The response rate was fairly uniform across schools, varying between 20% (shass) to 10% (Sloan). The student response rate was 12%, with instructors and faculty responding at a 16% rate. We summarize a few of the key results.

Overall Usage (Never/Occasionally/Often/Very Often). The figure below shows that ChatGPT continues to be the most widely used system, with 44% of respondents (46% of students and 35% of instructors) saying they used it “Often/Very Often” (green bars.) Overall, Sloan reported a much higher rate of usage of all tools than any other group.

: Table summary: Usage frequency of AI services in academic activities, where Copilot, Other, and Self-hosted models show the highest rates of frequent use. Specifically, 77 percent of Copilot users, 84 percent of those using Other services, and 90 percent of Self-hosted model users report the highest frequency category. In contrast, ChatGPT has a more distributed usage pattern, with only 21 percent in the highest frequency category and 35 percent in the second. Claude and Gemini show similar profiles, with 45 percent and 44 percent of their respective users in the highest frequency category.

Faculty/Instructor Usage (No Orange/Not Applicable/Yes Blue). We asked instructors how they used A.I; many use it for administrative tasks and helping prepare assignments or lecture notes, with much lower numbers using it for grading or individualized feedback.

Table summary: AI tool usage is highest for creating lecture notes at 61 percent and generating learning activities or assessments at 60 percent. Routine administrative tasks are used by 49 percent of respondents. In contrast, AI is rarely used for automating grading, giving feedback on student work, creating personalized learning plans, or providing virtual teaching assistance, with each of these tasks reporting usage rates of 7 percent or lower.

Student Usage. Students use A.I for many purposes, with large fractions reporting using it for coding, information search, writing revision and summarization. Fewer use it for whole-cloth generation of new text. As with overall usage, Sloan students use A.I for all purposes at a significantly higher rate than other groups.

।Image summary: A horizontal bar chart showing the percentage of responses for several unnamed tasks, all with a sample size of N>30. The percentages range from a high of 48% for the top task to a low of 2% for the bottom task, with other tasks falling at 45%, 42%, 40%, 20%, 15%, and 13%.

Table summary: Coding and programming assistance is the most common use case at 48 percent, followed closely by searching for information at 45 percent and revising writing at 42 percent. Summarizing information also sees high usage at 40 percent. Other tasks are used significantly less, with translating text at 20 percent, routine organizational tasks at 15 percent, and both generating multimedia content and generating text from scratch at 13 percent. These results represent all responses with a sample size greater than 30.

A.I Attitudes (Strongly Disagree Brown / Somewhat Disagree / Neutral / Somewhat Agree / Strongly Agree Dark Green). Respondents broadly feel that they need A.I skills and that A.I makes them more productive, but also that they are becoming more dependent on it. Responses are more mixed on whether it helps learning or whether it is desirable in class. Most people do not like to use it as a writing tool.

How much to do you agree or disagree with each of the following statements?...

: Table summary: Student perspectives on AI, where the strongest agreement is for the necessity of learning AI skills for future careers, with 45 percent reporting the highest level of agreement and 32 percent the second highest. There is also a high level of concern regarding AI dependency for coursework, with 36 percent agreeing most strongly and 32 percent agreeing. In contrast, there is low preference for AI writing tools over traditional methods, with 33 percent reporting the lowest level of agreement and only 6 percent reporting the highest.

Research Use (Never Acceptable Brown / Sometimes Acceptable / Unsure / Always Acceptable Purple). We asked what types of A.I use the community felt were appropriate. Most felt that spelling and grammar, research, and code generation were acceptable uses, while most felt that using it for peer review or drafting paragraphs from outlines or notes was unacceptable. Other uses had mixed views, reflecting the difficulty in creating a uniform policy for research usage.

Table summary: Perceived acceptable uses of AI in scientific or academic research, based on responses from over 30 participants. The most widely accepted uses are assisting with spelling, grammar, or language when writing a manuscript, with 68 percent agreement, and searching for relevant literature, with 58 percent agreement. Other tasks show a more divided or lower level of acceptance. For example, analyzing data and or modeling and editing images or figures are seen as acceptable by 28 percent and 19 percent of respondents, respectively. The least acceptable use is peer-reviewing anonymous manuscripts, with only 6 percent agreement.

Table summary: User concerns regarding AI use, categorized by the level of consideration given to specific issues including inaccuracy or hallucinations, societal impact, ethics of training models, and environmental impact. Across all categories, respondents are most likely to view inaccuracy and hallucinations as a major consideration, with 74 percent rating it as such. In contrast, environmental impact and the ethics of training models are viewed as major considerations by 32 percent and 34 percent of respondents, respectively. Societal impact is considered a major concern by 40 percent of users. For each category, the data shows a distribution across three levels of consideration: not a consideration, minor consideration, and major consideration, with the majority of responses for inaccuracy and hallucinations falling into the major consideration category.

Definition

Cognitive development: The progression of thinking, reasoning, problem-solving, and learning capabilities throughout an individual's life.

Approximately 10% of respondents (N=163) wrote in a concern for the "Other" response here. The most common concern (N=79) was the potential impact of A.I on learning and cognitive development. Respondents expressed concern that reliance on A.I could reduce critical thinking, writing ability, creativity, reasoning, and long-term skill development, or otherwise diminish the educational value of coursework. A second broad category (N=34) reflected more general philosophical or societal concerns about the role of A.I in education, including worries about the replacement of human expertise, the purpose of higher education, and the broader societal implications of A.I. Other concerns included the usefulness, reliability, or educational value of current A.I systems (N=20); the practices and business models of A.I companies (N=14); privacy, security, copyright, and intellectual property (N=10); academic integrity and authorship (N=7); the cost of A.I tools (N=5); and potential economic or labor-market impacts (N=4). Overall, although the concerns were diverse, they were dominated by questions about how A.I may affect student learning, cognitive development, and the educational mission of the Institute.

C.2 Spring 2026 Quality of Life Survey

The Spring 2026 campus-wide Quality of Life Survey also asked questions about generative A.I usage and attitudes. We also briefly summarize these results here. This survey was given to all members of the M.I.T community, with about 8200 respondents from the main M.I.T campus (Lincoln Labs was also surveyed but their data is omitted in the analysis below). For this analysis we group faculty and instructors together, postdocs and grad students together, and service and support staff together.

Frequency of A.I Tool Use. We asked respondents how frequently they used generative A.I tools. Overall about 40% of respondents said they used these tools often or very often, with 46% of undergrads and 60% of undergrads and graduate students reporting at least one次. Service and support staff had a markedly low level of use, with 53% reporting they never used A.I tools.

.Image summary: A stacked bar chart showing the frequency of generative AI tool use among various roles at MIT's main campus. Of the total 8,431 respondents, 21% never use these tools, 39% use them occasionally, 22% often, and 19% very often. Use varies by role: Service/Support staff report the lowest use, with 53% stating they never use AI; Graduate Students/Postdocs report the highest frequent use, with 33% using it very often and 27% often.

Generative A.I Attitudes: We asked respondents a series of questions about A.I attitudes: whether A.I made them more pessimistic or optimistic, whether it made them feel more replaceable or capable, whether it made their work better or worse, whether it made them more or less efficient, and whether A.I was unreliable or reliable.

Pessimistic vs Optimistic. Overall, only 23% of respondents are optimistic about generative A.I (rating a 4 or 5 on a pessimism to optimism scale). Instructors and grad students/postdocs report a slightly higher level of optimism.

How do you feel, generally, about generative A.I along these dimensions: Makes me pessimistic about the future (1): Makes me optimistic about the future (5) Main Campus only. Response range: 1 -greater than 5.

: Image summary: A stacked bar chart showing survey ratings from 1 to 5 across six different groups. For the Total group (n=8,245), 34% gave a rating of 3, 22% gave a 1, and 21% gave a 2. The Service/Support group (n=813) had the highest percentage of 1 ratings at 31%. The Grad Students/Postdocs group (n=2,348) had the highest percentage of 5 ratings at 11%.

Keepable / Capable. The community is largely divided on the replaceable/capable axis. Overall, 40% of respondents said they felt A.I made them more capable, and 27% said they felt it made them more replaceable. Postdocs, grad students, instructors, and admin staff reported that A.I made them feel more capable at a significantly higher rate than undergrads and service/support staff. Notably, undergrads reported feeling replaceable at a higher rate than capable (40% versus 34%).

How do you feel, generally, about generative A.I along these dimensions: Makes me feel replaceable (1): Makes me feel more capable (5)

Image summary: A stacked bar chart showing survey ratings from 1 to 5 across different academic roles. For the total population of 8,207, the distribution is 12% for 1, 15% for 2, 32% for 3, 27% for 4, and 13% for 5. By role, Service/Support (n=806) has the highest percentage of 1s at 21% and the highest percentage of 3s at 41%. Faculty/Instructors (n=701) have the highest percentage of 5s at 18%. Grad Students/Postdocs (n=2,344) have the highest percentage of 4s at 32%. Admin (n=2,572) and Undergrads (n=1,360) show intermediate distributions, with Undergrads having the highest percentage of 2s at 23%.

Worsens / Improves Work. Respondents reported that they felt A.I improved their work more than they felt it worsened their work (42% selecting a 4 or 5 on the work worse / work better axis vs 21% answering a 1 or 2). Differences between subgroups were not large, although undergrads reported a slightly higher rate of “worsens work”.

How do you feel, generally, about generative A.I along these dimensions: Makes the quality of my work worse (1): Makes the quality of my work better (5)

Main Campus only. Response range: 1 -greater than 5.

Image summary: A stacked horizontal bar chart showing survey responses on a scale of 1 to 5 across different roles, with a total sample size of 8,192. For the total population, the largest response group is 3 at 37%, followed by 4 at 30%, 2 at 11%, 5 at 12%, and 1 at 10%. By role, Service/Support (n=800) has the highest percentage for response 3 at 49%, while Grad Students/Postdocs (n=2,348) have the highest percentage for response 4 at 36%. Undergrads (n=1,357) show the highest percentage for response 2 at 18%.

Inefficient / Efficient: There was strong agreement that people felt A.I made them more efficient, with 60% of respondents selecting a 4 or 5 on this axis. The service / support group reported a lower rate of “makes my work more efficient”; otherwise differences between groups were small.

How do you feel, generally, about generative A.I along these dimensions: Makes my work less efficient (1): Makes my work more efficient (5)

Image summary: A stacked bar chart comparing responses across six groups based on a scale of 1 to 5. For the Total group (n=8,193), the largest proportions are 39% for 4 and 21% for 5. Among Grad Students/Postdocs (n=2,347), 48% responded 4 and 32% responded 5. Undergrads (n=1,360) show 49% responding 4 and 22% responding 5. Service/Support staff (n=804) have the highest proportion of 3s at 47%, with 20% responding 4 and 9% responding 5. Faculty/Instructors (n=700) show 32% responding 3, 36% responding 4, and 20% responding 5. Admin staff (n=2,559) show 37% responding 3, 34% responding 4, and 16% responding 5.

Unreliable / Reliable: Significantly more respondents felt that A.I was unpredictable and unreliable than felt it was trustworthy and reliable, with 39% selecting a 1 or 2 vs 21% selecting a 4 or 5 on this axis. Again the differences between subgroups were small.

How do you feel, generally, about generative A.I along these dimensions: Current systems are unpredictable and unreliable (1): Current systems are trustworthy and reliable (5) Main Campus only. Response range: 1 -greater than 5.

Image summary: A stacked horizontal bar chart showing responses on a 5-point scale across different academic roles. For the total group of 8,250 participants, the highest response percentage is 38% for rating 3, followed by 23% for rating 2, 20% for rating 4, and 16% for rating 1. Among specific roles, Service/Support staff have the highest percentage of rating 1 at 27% and rating 3 at 42%, while Admin personnel show the highest percentage of rating 3 at 44%.

M.I.T Guidance. We also asked respondents whether they thought M.I.T had given them adequate guidance about A.I use. Undergrads largely reported that they had received such guidance, with 72% reporting they somewhat or strongly agreed, with other groups reporting much lower levels of adequate support.

M.I.T has provided me adequate guidance on how and when to use generative A.I in my work or coursework at M.I.T.

Main Campus only. Response range: Strongly disagree -> Strongly agree.

Image summary: A stacked bar chart showing agreement levels across different professional roles for a total sample of 8,399 people. The categories range from Strongly disagree to Strongly agree. For the Total group, 44% agree (30% somewhat, 14% strongly), 24% are neutral, and 31% disagree (18% somewhat, 13% strongly). Among the roles, Undergrads show the highest agreement at 72% total (47% somewhat, 25% strongly), while Service/Support show the highest disagreement at 41% total (17% somewhat, 24% strongly). Faculty/Instructors have the highest neutral response at 34%.

$ ^{29} $https://qol.mit.edu/

C.3 Fall 2025 Tech Survey

In the Fall of 2025, the Tech (M.I.T's student-led newspaper) also conducted a survey of student attitudes and usage of A.I. The results are described in detail in a recent tech article, which we summarize briefly here. The survey received responses from 1,002 affiliates, including 659 under- graduates, 248 graduate students and postdocs, 18 faculty, and 72 staff. Among undergraduates, 46% reported using L.L.M's daily and 30% several times per week; among graduate students and postdocs, the corresponding figures were 46% and 35%. The survey found that the most common use among undergraduates was explaining course material (over 80%), followed by coding assistance (around 70%), coursework completion (over 50%), and summarizing papers (over 50%). Among graduate students and postdocs, the most common use was programming assistance (88%), followed by brainstorming ideas (55%), summarizing papers (52%), and writing essays (48%). The survey also found that 35% of undergraduates reported that L.L.M's save them "a lot" of time and more than 45% reported that they save "a bit" of time. The survey found that 90% of undergraduates were somewhat or very concerned about overreliance on L.L.M's, including 67% who were very concerned. In addition, 45% were very concerned about inaccurate or misleading outputs.

The survey found that 75% of undergraduates believed faculty expectations regarding A.I use were clear, while 50% believed decisions about A.I use in coursework should be left to individual instructors and departments. Although 70% agreed that A.I proficiency will be important in their careers, only 25% believed M.I.T is preparing students to use A.I professionally.

$ ^{30} $https://thetech.com/2025/11/25/llm-survey-results

D. Additional Resources

See aihub dot mit dot edu for more resources about A.I in education and research and how we are beginning to implement the findings of this report.