Skeptical, Ethical, and Ambitious

Getting AI “right” in schools has turned out to be much, much harder than most people anticipated - it doesn’t help that what “right” even means is a huge part of the problem. I have lived through several technology transitions in thirty years of teaching - the internet in the late 1990s, social media and smartphones in the late aughts, the Google ecosystem in the 2010s, and, most recently, the pandemic-era pivot to distance learning coupled with the proliferation of online learning platforms over the past five years. None of them produced anything close to the current level of polarization around AI. The poles in the public debate surrounding it (ban / embrace) often impede the likelihood of having fruitful conversations in schools. The most incendiary pieces on both sides, and the inevitable rounds of hype and dismissal, frequently crowd out reasonable attempts at nuance. As a result, most educators have found themselves caught in a crossfire they are remarkably ill-equipped to navigate. It’s no wonder academic leaders are so paralyzed and confused.

Which is why a recent letter from the President of the University of Chicago caught my attention. The most thoughtful passages read as if at least one institution is finally getting the AI moment right, or as right as the moment allows. I found myself wondering if there were any lessons other schools might learn from their approach.

On June 2, President Paul Alivisatos announced that the University of Chicago has partnered with Anthropic to roll out Claude Enterprise to faculty and staff in July, and to students before fall. The partnership headline got much of the attention (and blowback), but the letter itself is worth parsing carefully.

Alivisatos asks the community to be “skeptical, ethical, and ambitious when it comes to AI.” Those three words rarely appear together and almost certainly weren’t the result of an AI “rule of three” application. Each was carefully chosen.

The first is a reflexive stance to resist hype. The second recognizes that AI’s impact extends far beyond its immediate classroom uses. The third grants permission to use the technology in service of the school’s mission. Taken together, the triad offers a framework more schools could emulate. It cautiously captures the genuine excitement around the technology without surrendering to it, and it acknowledges the legitimate concerns without retreating into prohibition - a position shown to be untenable over the past three years.

The letter does more than just take a clear stance, though it does that effectively. It also commits to a set of practical decisions. The newest versions of the tools will be made available with university data protection which means employees currently using private accounts no longer need to. Extending Enterprise to all employees - faculty, staff, and operations - rather than limiting it to academic use is a structural commitment few institutions have made publicly, an implicit recognition that frontier-tier access matters across university infrastructure, not just in the classroom.1

Crucially, the letter grounds AI engagement in what Alivisatos calls the “habits of mind and standards of judgment that define this university.” AI integration is positioned not as a departure from UChicago’s existing intellectual mission but as a continuation of it. The letter also explicitly rejects the “AI is always better” framing: “In some cases, it is clear that AI tools can be brought to bear with the desired results. In others, we will do better by retaining our previous approaches or even developing new policies to expressly protect against misuse.” This line should assure critics who fear AI is going to infuse every aspect of the student learning experience.

Speaking directly to students, Alivisatos writes that:

… the University has a duty of care to ensure that the education offered to you is responsive to these technological developments by teaching you how to think with machines, how to think without them, and how to think about them. (emphasis added).

That single sentence contains more practical wisdom than the densest AI policy memos.

Faculty resources linked in the letter are home-grown, curated, specific, and authentically the product of sustained faculty work. They represent the kind of investment of faculty time across multiple departments that does not materialize in a few months. Two new faculty committees - one on AI and academic life, chaired by William Hubbard of Law and Economics, and an advisory committee on AI and education, chaired by Emily Lynn Osborn of History - signal that institutional commitment is ongoing, not a one-off announcement.

Chicago’s position didn’t appear overnight. It is the visible output of a faculty process that has been running in the open since early 2025, when Provost Katherine Baicker convened an AI and Education Working Group. That group’s report gave the University its governing phrase: be skeptical, ethical, and ambitious. I haven’t been able to read the report itself (it sits behind a university login), but Alivisatos has described its thrust the same way in every public message I can find, from an October note to faculty through this letter.

The June announcement is the logical next step and culmination of that process. UChicago has run its own tool, PhoenixAI, a custom platform built on OpenAI’s models, available to faculty and students for over a year. Its Becker Friedman Institute has put Anthropic’s models in the hands of its economists. None of this was improvised. Most institutions cannot point to the same trajectory.

The UChicago letter contrasts nicely with what happened at California State University last spring. A year ago, the CSU system signed the largest single-institution ChatGPT deployment in the world, a $17 million, 18-month deal with OpenAI rolled out across 22 campuses. CSU renewed in May 2026 for $13 million annually over three years. The New York Times Magazine recently characterized the result as “chaos.”

The UChicago decision, by contrast, emerged from a process the community could see coming. While clearly there is some disagreement within the UChicago community2 - no decision in the current climate around AI is immune - my admiration stems less from the decision3 itself than from the principles and process that produced it. The financial terms of the deal have not been disclosed, which is a significant criticism the partnership has invited. The framework, however, is public. A school community deserves open, defensible principles - and on that score, UChicago has demonstrated intellectual seriousness that suggests they can change course if the rollout ends up doing more harm than good.

When we study the Supreme Court in my Government and Politics class, I tell my students that no case that makes it to the Supreme Court is “easy.” Cases that survive the appeals process and are granted certiorari pose hard legal questions. If the decision were obvious, it wouldn’t be in front of the highest court.

The same logic applies to school leadership decisions on AI. One reason school leaders have been reluctant to take a strong stand is that the landscape is incredibly fraught. The technology is improving faster than anyone can follow, threatens the essence of what schools have been doing for decades, and allows students to complete work with little to no evidence of learning. It is forcing a logistical, pedagogical, and existential crisis on a scale schools have rarely had to handle, and it is not going away. At the same time, as Alivisatos notes, “faculty and researchers at all levels are using AI systems in astonishing and inspiring ways.” The juxtaposition of these two realities is what makes determining where AI might have a place on college campuses so difficult.

Alivisatos’s decision will appeal to some and upset others. That comes with the territory of being a school leader. Many schools will not arrive at the same place as UChicago. The relevant question is whether the decision was reached through reasoning that critics can actually engage with. In this case, it was. Critics need to specify exactly what was wrong with the reasoning, not just object to the outcome. Alivisatos’s language is deliberative and forward-looking at the same time, which is precisely the posture schools need to be in right now.

Schools downstream of higher education - especially the K-12 institutions I work with directly - now have at least one model of what thoughtful engagement with the AI question looks like. Schools need principles, not just policies. Policies are rules; principles are how those rules adapt.

At the end of the day, any letter to the community about its AI principles is just words. But words matter when they are coupled with significant institutional decisions. The choice to partner with Anthropic (or OpenAI or Google, for that matter) is a major commitment. It’s incumbent on any school, not just UChicago, to follow up on its words with ongoing adjustments, calibrations, and actions across years, regardless of which direction they decide to move in. Many schools still have not chosen to do this kind of work. The few that do are starting to draft Phase II of the AI conversation in education.

Beyond this newsletter, I work directly with schools, educators, and other institutions about pedagogical questions raised by AI. Take a look at my website and reach out - I’d love to hear what you’re working on.