Among its key findings, the report revealed that while most students report using generative AI in the course of their studies, many of those same students worry that it is negatively affecting their long-term cognition. Faculty members shared these concerns, and while many faculty have begun using AI in some capacity in teaching or research, many course syllabi do not adequately establish clear expectations or limitations on AI use.

The GAITL committee’s immediate recommendations included publishing guidelines that help to clarify expectations around AI. Intermediate and longer-term recommendations include updating academic codes to address generative AI, and eventually partnering with peer institutions to set standards around its use in teaching and learning. In August 2026, an expanded committee, called GAITL Phase 2, shared sample generative AI syllabus statements to serve as a resource for faculty in their courses, and Doyle has charged the group with engaging with the campus community about the report’s longer-term recommendations.

In an interview, Doyle and Littman discussed the report, its development process, and next steps in Brown’s effort to incorporate AI in teaching and learning in a way that harnesses its positive potential and mitigates risks to academic integrity and the ability of faculty to assess learning and understanding. 

Q: For universities, the proliferation of generative AI tools has raised clear challenges and presented new opportunities. How has Brown gone about confronting and considering those?

Doyle: Our approach has been to make sure we are taking this on in a way that is consistent with our institutional values. Brown has a long tradition of approaching challenges this way. I am going to take us back in time a bit. There is an article in the Brown Alumni Magazine from 43 years ago about how Brown should incorporate computers in education. It speaks to some of the same challenges and opportunities we face today, and there is a quote from an associate provost at the time that just screamed out at me: “If you accept that fact and realize that it has the potential for major social change — both positive and negative — and if you realize that many schools with different goals may be shaping the technology, you come to the conclusion that we have the opportunity to shape the new technology, too, and to do it in a way that is really appropriate to Brown.”

For me, there is a strong resonance with what we are seeing with AI. We cannot sit and wait for this to happen to us. Much like our very proactive mentality in the ’80s when it came to the computer and education, I think we are in a moment where Brown has a chance to seize this opportunity and leverage the Brown DNA. And that has been our approach. We created the associate provost for AI role, and we stood up the Generative AI in Teaching and Learning Committee so that we could engage the entire campus in this process of shaping an approach to AI that is consistent with our institutional values. And now we have the GAITL report that charts a way forward. Q: Could you talk a little about the report and how it came together?

Q: Could you talk a little about the report and how it came together?

Littman: We formed the committee in spring of 2025 with representatives from across campus with the goal to try to understand: Where were we in terms of how AI was influencing teaching and learning? What is happening on campus? What is happening on other campuses that we should know about? What is the research around this starting to look like? At one point, one of the committee members said, “Well, how should we be thinking of this in terms of the Open Curriculum?” That was a big frame shift. The whole committee clicked into place at that point. There was a realization that we’re not trying to solve this problem at a global scale. We’re trying to figure out what this means for Brown and how we can shape that future.

Q: I want to dig in a bit on the Open Curriculum. How does it influence Brown’s approach to understanding AI and incorporating it into how we teach and learn?

Doyle: We have great examples from across campus of faculty creating opportunities for students to explore AI in many different ways in the classroom, in the lab and out in the world. That is not just in computer science, engineering and applied math where Brown faculty are at the cutting edge of developing new AI tools that solve all kinds of real-world problems. We have also had classes that examine AI in the humanities, in law and public policy, education, healthcare and other areas. Those opportunities are there, and the Open Curriculum enables students to take full advantage of them. 

In terms of top-down policy, however, the Open Curriculum is a bit of a double-edged sword when it comes to AI. On one hand, it offers students real flexibility to take ownership of how deeply they immerse themselves in AI. If we are able to provide clarity about the content of courses up front, students can be intentional about leaning in different directions in terms of AI engagement. But it also presents a challenge. By design, and to promote intellectual curiosity, we do not have narrow lanes that students need to pass through as they chart their intellectual paths, so it is challenging to create checkpoints in students’ exposure to AI. 

I think we do have an opportunity to create areas in the curriculum that emphasize intrinsically human skills and veer away from reliance on AI tools. And in other areas, instructors may employ a deeper immersion into developing, using and evaluating AI tools. And there is a full spectrum of possibilities in between. The key is developing markers or indicators that make expectations surrounding AI clear to everyone up front. That is something in the report that grew out of the discussions we have been having over the past year.

Littman: I agree. I feel like the openness, the flexibility and the exploratory nature of the Brown education is kind of perfect for the kind of topics that we’re now dealing with. For example, we’re trying to re-envision education at some level — how we engage students and how we assess them, what material is central and what is just interesting — and it’s a long-standing cultural perspective at Brown for the students to be involved in that and not just subject to it.

Q: The report posits that traditional approaches to ensuring academic integrity may not fully address the realities of generative AI. What is Brown doing about concerns over AI cheating? How should Brown and other schools rethink integrity without creating an atmosphere of mistrust?

Littman: This is one of the main topics the committee focused on. Among the top concerns shared by Brown faculty and students regarding the increased use of generative AI were that it would reduce students’ long-term learning, have negative consequences for cognition and encourage student cheating. There are lots of different perspectives at Brown and nationally on how to confront these challenges, and lots of disagreement, but one thing we as educators all agree on is that the point of teaching is for people to learn, and to the extent that certain uses of the technology are undermining that, we have to find a way to fix it.

And students are part of this conversation, too. I have spoken to many students who are concerned they’ll feel forced into a position of using a tool that may be actively harming their cognition. We have to recognize that we are trying to sort of reimagine how higher education works to some degree. That has to be a broad conversation. There’s nothing we can just implement by fiat. We have to talk to each other.

In the meantime, Brown’s faculty-led Standing Committee on the Academic Code continues to do its work in assessing and addressing reports they receive about cheating or other code violations. One of the immediate-term recommendations in the report is that the committee should take documented AI rules — whether they come from the University level, a class syllabus or simply written rules for a specific assignment — into account when assessing concerns about cheating, plagiarism or other potential instances of academic dishonesty that violate our academic standards. 

Doyle: It is important to emphasize that we are not talking about changing our mission or our core institutional values when it comes to safeguarding academic integrity. We hold those dear: Free inquiry, openness, responsibility to the academic community. I think we all have to share some of the burden of responsibility to think differently about how we do things like assessment — this extends to testing and exams, and grading papers and lab work. I can say, as an engineering professor, that the advent of calculators and laptops forced us to rethink how we did certain assessment modalities for engineering students. I think that is germane here. The reality, as Michael likes to say, is that students have the equivalent of the answer key in an envelope in their desk. Faculty really do have to think, “What are the right modalities to test understanding and learning?”

Q: What were some of the most interesting and important things you learned during the course of the committee’s work so far?

Doyle: One of the things I found fascinating was the deep dive they did on the course syllabi outlining goals and expectations for courses across Brown. They pulled out almost 3,000 syllabi and they applied locally run large language models to study them. And one of the conclusions that came from that is that our faculty do not have clear policy statements in their syllabi. But they have to talk about this at the beginning of class. General statements about academic integrity that do not cover AI need revision in this moment. And that is something I hope we can very quickly change, and provide support for that transition.

Littman: Yes, that’s an important point. More than half of the syllabi we looked at did not have anything to say about AI use. Since it’s really hard for there to be any one campus-wide policy on the question of what AI use is appropriate in what class, it is very important for faculty to address it up front in their individual courses. It is really difficult to hold students accountable to an academic standard for use of AI when no standard is articulated. Other institutions are grappling with the same issues.

Q: What sort of resources do you anticipate being available to faculty who aren’t sure what that syllabus should look like?

Littman: We’ve put together a GAITL Phase 2 committee, which is working to act on some of the recommendations made in the original GAITL Committee’s report, including resources for faculty and guidance about what tensions faculty should be thinking about as they design their course syllabi. If we’re going to approach the incorporation of AI in a way that maintains academic freedom — which is an imperative — then it really is up to individual faculty members to grapple with this topic, with support and resources from Brown.

For example, the web page for our Sheridan Center for Teaching and Learning now includes a handful of exemplary AI policies that are in syllabi on campus right now. Some examples show how to make it very clear that AI is not appropriate at all for this class. There are also examples that are much more AI-forward — all the way through classes that explicitly explore AI tools. The other thing that we’re planning to release is a resource to help faculty think about use cases and how students might be using AI.

Doyle: If I could add one thing that we want to underscore, it is that the faculty have full agency to establish their policy. As Michael said, academic freedom is a critical consideration here. 

Q: Brown has been around for 262 years, and its leaders and faculty have managed many other emergent factors that influenced education. Relatively speaking, how big a disruptor is generative AI?

Doyle: I think when you are in the eye of the storm, there is a sense of impact that I might respectfully say could be exaggerated. We have lived through relevant examples: The advent of calculators, computers and the internet come to mind. There was resistance to embracing those things in our teaching and learning methods, but they did not completely ruin fields or destroy pedagogical methodologies as some said they would. Of course, AI is also different from those other tools in ways that advocate for caution. It has a breadth of impact that is simultaneously affecting research, operations and teaching in ways that are unprecedented. But as you note, we have been here for 262 years, and I fully expect we will be here for the decades and centuries to follow.