In this reading
Starting now, the MIT community must embark on a set of changes to make our educational approaches more AI-aware.
An independent, AI-assisted editorial reading. Quote blocks are verbatim excerpts from MIT's report. The framing and suggested actions are our own; the report's recommendations are not presented as a completed rollout.
Why the work must begin now
We are not waiting for AI to enter education. We are deciding how to respond to changes already affecting the work, the evidence of learning, and the relationships through which education happens.
AI 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.
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.
AI is changing what students need to know and know how to do.
The urgency is not simply that assignments are easier to complete. It is that completing them may no longer mean what we assumed it meant. Waiting for certainty does not preserve the old conditions; it leaves those changing conditions to shape the education we provide.
But as a community, what should worry us most is that many uses of AI deprive students of the opportunity to learn.
Yet uncertainty can't be an excuse for inaction.
Be clear about what we are acting to protect
The point is not to defend every familiar assignment, nor to make AI adoption an achievement in itself. It is to preserve what an education gives a person: understanding they can exercise, judgment they can trust, and the ability to contribute to a world shared with others.
The most important product of their education is not a GPA or a diploma but themselves: their personal growth and intellectual maturity and the development of their own imagination, insight, and judgment.
AI should be used to augment and enhance curiosity, creativity, and learning, not automate them.
That purpose changes our decisions. A research position is not only a way to get research done. A difficult assignment is not only a deliverable. A conversation with a teacher or peer is not merely a slower route to an answer.
While undergraduates can provide faculty with useful research labor, that is not the point of the program. It exists to educate.
This is the core of MIT: In the spirit of "Mind and Hand," we work together on real projects that require deep thought and careful, rigorous work.
Preserve the purpose. Be willing to change the form.
Redesign learning with intention
Start with what students need to understand and be able to do. Then align the practice, assessment, and AI policy with that goal. A rule about a tool cannot substitute for a reason to do the work.
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 AI.
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.
When AI makes it possible to offload the cognitive work of learning, how can we assess what students actually know and understand?
The report's eight principles give this redesign its direction:
- Be humble. Expect to revise what we learn from practice.
- Be bold. Begin before every uncertainty is resolved.
- Put humanity front and center. Judge the change by what it does for people.
- Lean into learning. Protect productive struggle and explain its value.
- Teach with intentionality. Let the learning purpose govern the design.
- No one size fits all. Match the approach to the subject and the learner.
- Augmentation not automation. Develop the learner's capacities, not only their outputs.
- Think beyond the classroom and the campus. Prepare people for responsibilities beyond the course.
The principle names are MIT's; the short explanations are our editorial reading. 2.1 2.4 2.5 2.6 2.8
That need not mean incorporating AI into everything we do, or even most things.
Use the opportunity to make education more ambitious
Protecting learning is not a call to retreat. AI can open new ways to practise, investigate, and create. The question is what those possibilities allow the learner to do next.
AI 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.
Architecture students are using AI to experiment with new ways to visualize and rapidly test their ideas, beyond what's possible with traditional representational skills.
The report gives a concrete example of assistance that increased practice rather than replaced it:
For instance, an instructor reported to us that, for students learning to serve as mediators, providing personalized, course-specific AI 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.
This is a reported classroom experience, not a guarantee about AI coaches. Its value is the design question it reveals: does the help return the student to the work with greater ability, or make doing the work unnecessary?
Use new capabilities to expand what students can attempt. Keep the understanding, judgment, and responsibility that the attempt should develop.
Change the work in front of us
The report asks for changes to learning goals, assessment, policy, and shared experience. These are not separate conversations to postpone indefinitely. They meet in the next course, assignment, project, and interaction we design.
Make assessment more valuable for learning
We urge instructors to consider forms of assessment that are less vulnerable to AI, and more valuable for learning, such as oral exams, semester portfolios, and out-of-class assignments paired with in-class conversations.
Moving everything into timed evaluation is not a complete answer. The report warns that it can reduce the incentive for sustained, difficult work and the time available for thought. Protect evidence of individual understanding without reducing the education to what is easiest to police. 3.1.2
BeginTake one substantial assignment. Decide what a conversation, demonstration, portfolio, or project check-in would reveal that the final submission alone cannot.
Make the learning reason for each policy clear
Therefore, instructors, and perhaps departments, should make sure that every MIT subject has a clear policy about the use of generative AI, posted prominently in the syllabus and on the course website.
When instructors make clear why AI is permitted, limited, or required, students are more likely to understand the learning that's being protected or developed.
A common framework does not require an identical rule in every course. It requires intelligible choices. A beginning student building foundations and a researcher working within established expertise may need different boundaries. 2.6
| Option | Original policy wording |
|---|---|
| Unrestricted GenAI use | "Students may use any GenAI system for any purpose on assignments in this course." 3.1.8 |
| Limited GenAI use: support tool only | "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." 3.1.8 |
| Required GenAI use | "Students are required to use GenAI to complete the assignment in the manner specified." 3.1.8 |
| GenAI use strictly prohibited | "GenAI may not be used in any form for the course." 3.1.8 |
These excerpts are not complete policies. Required use needs a specified tool or class of tools, workflow, and documentation; it can coexist with prohibited components. The report also warns that a blanket prohibition on out-of-class use is difficult to enforce and can create risks of missed violations and false accusations. 3.1.8
BeginState what is allowed, what is not, which stages the rule applies to, and the learning reason. Check whether students understand the explanation, rather than assuming that publishing it is enough.
Keep clarity and preparation distinct
In the fall 2025 Tech Survey, more than two-thirds of students who responded felt that AI would be important in their careers, yet only 25% felt that MIT was adequately preparing them to use AI.
The web report also records that students find AI guidance confusing. Knowing a rule, understanding its purpose, and being able to use AI well are different questions. Find out which problem needs addressing. 3.2.4
Make expectations reciprocal
If instructors plan to present students with content substantially generated by AI, or to use AI for some aspect of evaluation, grading, or feedback, we strongly recommend that they be transparent with their students about how and why AI is being used.
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.
BeginExplain how and why instructors use AI alongside the expectations placed on students. Decide where human feedback, judgment, and availability remain essential.
The committee makes its own AI contribution visible
MIT's published web report asks instructors to be transparent about how and why they use AI. That is the standard worth copying: explain the contribution, rather than merely announcing that AI was used. 3.2.3
Make shared learning part of the design
Every subject should include a regular in-person social component (not just sitting in lecture or recitation and quietly taking notes).
Group projects with weekly staff check-ins and deliverables that assess both individual and collaborative contributions.
Feedback discussions structured around a class rubric.
Shared work needs suitable space, facilitation, and a way for each person to contribute. The report also proposes guided problem-solving, facilitated discussions, and staffed collaborative spaces across disciplines. 3.1.4 3.1.8
BeginBuild a recurring interaction in which learners explain, question, solve, or give feedback to one another. Make its educational purpose explicit.
Build shared life beyond the classroom
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.
"Tech Free Times" (periods during which MIT would not schedule classes, office hours, or meetings and instead support activities centered on in-person connection).
The report also proposes broader participation in MIT Reads, celebrations of human skills and accomplishments, and expanded first-year learning communities. These offer different ways to create recurring opportunities for connection; they are not interchangeable events or guarantees of belonging. 3.2.2
Teach the judgment that responsible use requires
To use AI 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 AI at all.
AI 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 AI.
BeginGive learners practice verifying an output, explaining AI's contribution, and recognising when not to use it. Bring this into the discipline, not only a general introduction to tools.
Research requires its own explicit standards: the report calls for AI-use statements in theses, human responsibility for accuracy, and attention to the rules of journals and conferences. 3.2.5
Give people the means to act together
An expectation to change is not the same as the capacity to change. An educator can understand the problem and still lack the time, expertise, or support to redesign a course. Institutions must make the work possible.
If we expect instructors to make more than incremental improvements, we need to offer them guidance, direct support, and community.
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 AI.
Create regular opportunities to learn from the work
Holding regular "lunch-and-learn" style seminars where instructors can hear from colleagues about how they are using and adapting to AI.
Creating online and in-person training on the use of AI for teaching and administrative tasks, such as how to develop animations and simulations, use coding tools, build task-tracking systems, and so on.
Bring real assignments, tools, difficulties, and classroom experiences into these conversations. The report calls for ways to share what people are trying and learning, not simply another generic training exercise. 3.3.5 4
Fund the changes, including deliberately AI-free learning
We recommend that MIT and departments/schools/the college create an AI Pilot Fund that instructors can apply to for resources (such as AI credits, TAs, UROPs, and summer support) to do AI-enabled projects, explore the impact of AI on pedagogy, and create deliberately AI-free experiences.
That last detail matters. The objective is better education, not more AI use for its own sake.
Give the continuing work clear owners
To put our recommendations into practice, MIT should establish an ongoing committee.
The report also proposes local AI Leads to coordinate policies and course adaptation, and AI Fellows and an implementation team to support instructors, build shared tools, and study their use. Strategy, local judgment, implementation, and resources are different responsibilities. A workshop cannot carry them all. 3.3.1
BeginIdentify who will lead a change, who will help carry it out, what resources it requires, and how colleagues will learn from the results.
Learn from the changes, and protect people while making them
Acting now does not mean acting without care. Build the safeguards and the learning process into the work from the beginning. Otherwise we may create a system that is easier to operate and harder to trust.
MIT should begin tracking metrics around AI use, campus engagement, student satisfaction, and post-graduation feedback.
We must deepen our scientific understanding of how AI affects learning and make sure our educational practices reflect that evolving knowledge.
Track more than activity. Ask what learners can now understand or do, what relationships are strengthened or weakened, and what unwanted effects are appearing. That is our application of the report's call for continued evaluation, not a new MIT-prescribed metric.
Keep these responsibilities visible:
- Evidence and integrity: do not rely on AI detectors. Distinguish process evidence from proof of misconduct and make disciplinary standards clear. 3.1.9
- Privacy: establish the scope of classroom logging, who may access it, how it is retained, and how sensitive or distress-related material is handled. 3.3.9
- Fair access and model choice: check whether available tools and computing resources meet course needs; protect sensitive information without committing to one commercial ecosystem. 3.3.7 3.3.9
- Wellbeing and legitimate reluctance: recognise emerging problems, direct students to appropriate support, and consider lower-AI pathways where the subject allows them. 3.1.10 3.2.2 3.2.5
- Ethical and material costs: address authorship, training-data concerns, bias, and environmental and financial impacts. Do not mistake availability for an absence of consequences. 3.2.4 3.2.5 3.3.10
There are many unknowns about how best to integrate AI into the curriculum and how to design experiences to avoid its pitfalls.
BeginAgree what evidence will help you retain, revise, or stop an approach. Make room for findings that challenge the result you hoped to see.
Some changes take time. The beginning does not have to wait.
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.
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.
This is the distinction that makes the report a call to action. It does not ask us to finish an institutional transformation immediately. It asks us to begin the work that is necessary, including the work that will take time.
For a teaching team, begin with a learning goal, an assessment, and a clear explanation of AI use. For a department, coordinate the policies and support an experiment. For institutional leadership, provide the people, resources, access, and governance that let those changes endure.
The work will require thoughtful commitment and focused effort from the entire MIT community.
The goal: to preserve and enhance the distinctive transformative power of an MIT education. It is not too much to say that MIT's mission depends on it.
It was clear to all of us on the committee that, although our report is finished, MIT's work on this subject has only just begun.
The next step is not another declaration that education must change. It is a decision about the work we are responsible for, made with the people who can carry it forward.
There is a lot to do.
The full programme of work: all 26 recommendations
The main narrative above groups the work around action. This index preserves the complete numbered recommendation set, with brief editorial explanations and links back to MIT's report. These are proposals, not a completed implementation.
Adapt educational processes for an AI-aware world
| MIT recommendation | What the work entails |
|---|---|
| 3.1.1. Revisit course goals | Reconsider what students should know and do before deciding whether a familiar assignment or AI restriction still serves them. 3.1.1 |
| 3.1.2. Ensure durable learning through new course policies, structures, and forms of assessment | Use oral exams, portfolios, and conversations around substantial work. Avoid solving authorship at the cost of time for deep thought. 3.1.2 |
| 3.1.3. Emphasize experiential and project-based learning | Expand the scope and real-world application of projects. Preserve the conceptual skills and judgment needed to direct the tools. 3.1.2 |
| 3.1.4. Build structured in-person social learning into subjects | Design guided problem-solving, peer feedback, and group projects that reveal individual as well as collective contributions. 3.1.4 |
| 3.1.5. Preserve and expand out-of-class research and career experiences | Protect research apprenticeships, mentorship, and career exploration. Explore broader access, co-ops, and experiences beyond the lab. 3.1.5 |
| 3.1.6. Reconsider grades and incentives | Explore mastery and competency-based approaches and portfolio evidence. Do not intensify the incentive to cut corners by rationing top grades. 3.1.6 |
| 3.1.7. Expand in-person spaces for labs and in-person evaluation | Provide and staff collaborative spaces, including appropriate analog or AI-limited environments, across disciplines. 3.1.6 |
| 3.1.8. Provide AI use policies, with justification | Use a consistent format and a shared policy menu, with choices tied to course goals and explained to students. 3.1.8 |
| 3.1.9. Exercise caution with AI detectors and online exam platforms | Do not rely on AI detectors. Treat version history as process evidence, not automatic proof; clarify disciplinary evidence standards. 3.1.9 |
| 3.1.10. Support responsible experimentation in the curriculum | Allow coordinated AI-light and AI-heavy experiments, evaluate results, and remove administrative barriers without losing curricular coherence. 3.1.10 |
Center people, community, and the residential experience
| MIT recommendation | What the work entails |
|---|---|
| 3.2.1. Define and communicate the value of residential education | Explain why presence, shared work, and intellectual community matter. Design AI use to strengthen those relationships. 3.2.1 |
| 3.2.2. Strengthen social connection and personal wellbeing | Build recurring shared experiences, expand learning communities, and prepare staff to recognise problems and refer students to appropriate support. 3.2.2 |
| 3.2.3. Encourage instructor disclosure around their own AI use | Explain how and why instructors use AI in teaching, assessment, and feedback. Keep human effort and responsibility visible. 3.2.2 |
| 3.2.4. Teach effective, responsible, and ethical use of AI | Build learning from orientation through discipline-specific practice, including verification, disclosure, limits, and ethical judgment. 3.2.4 |
| 3.2.5. Recognize and mitigate negative impacts of AI | Take objections about training data, corporate power, and social effects seriously. Support research and policies that address harms. 3.2.5 |
| 3.2.6. Acknowledge AI use in theses and other research work | State how AI contributed, check journal and conference rules, and keep a human responsible for accuracy. Do not list AI as a co-author. 3.2.5 |
Build processes, teams, and tools for continuous improvement
| MIT recommendation | What the work entails |
|---|---|
| 3.3.1. Establish an ongoing AI and education committee | Give strategy, monitoring, evaluation, and policy coordination a continuing owner rather than treating the report as the end of the work. 3.3.1 |
| 3.3.2. Create school/college- or department-level AI Leads | Coordinate local policies, course adaptation, and new assessment models at a level close enough to understand the discipline. 3.3.1 |
| 3.3.3. Fund AI Fellows and an AI Implementation Team | Provide technology and learning-science expertise for individual instructors, shared tools, and observation of actual use and effects. 3.3.1 |
| 3.3.4. Create an AI Pilot Fund | Resource AI-enabled projects, pedagogical research, and deliberately AI-free experiences, including staff time and teaching support. 3.3.5 |
| 3.3.5. Provide ongoing training and instructor support | Build regular peer exchange and practical training for teaching and administration, rather than a generic training exercise. 3.3.5 |
| 3.3.6. Develop metrics | Track AI use, engagement, student satisfaction, and post-graduation feedback. Use findings to revise practice. 3.3.6 |
| 3.3.7. Ensure equitable technology access | Check whether shared access meets actual course needs. Account for model differences, coding-tool access, and required computing resources. 3.3.7 |
| 3.3.8. Protect sensitive data and preserve model choice | Protect private information while avoiding dependence on one commercial ecosystem. Consider appropriate local and open-model access. 3.3.7 |
| 3.3.9. Establish privacy, logging, and auditing policies | Resolve who can see what, retention, disclosure, and responses to sensitive or distress-related interactions openly and transparently. 3.3.9 |
| 3.3.10. Monitor AI costs and environmental impact | Make financial and environmental costs more visible, audit institutional use where possible, and support lower-impact approaches. 3.3.10 |