Weekly Message from HOS 2026/06/01-2026/06/05
Dear Keystonians,
For decades, great educators have pushed for change. Now, AI is finally forcing it.
Princeton and Stanford Universities recently did away with their academic honor codes, which were over a hundred years old. The reason for the change is generative AI, which has made it impossible for professors to have confidence that assignments were written by their students alone. For the first time in over a century, these institutions will require proctored exams.
While I was in the United States for my daughter’s graduation, I saw a colleague who is long past retirement age but continues to teach because he “can’t imagine what else he would do.”
He is a social studies teacher who specializes in government and economics and has spent over a decade leading trips of students and teachers to Bhutan. When I saw him on this visit, he said he was finally ready to retire. When he told me, he shrugged and said, “Because of AI.”
These things feel connected to me. The retirement of someone I thought would teach forever. The fall of century-old honor codes at prestigious universities. They are signals of an old system finally giving way—a system that great educators have been trying to move past for decades.
Do we lament that our outdated assessment models are now being challenged? Or do we turn to models that are tried and true, and confidently move forward?
The skills, competencies, and habits of mind that our students need to thrive are clear. The question for educators now is how do we teach those things, and how do we know they are being learned? I have written about the three central questions for educational design in this space before:
What is worth learning?
How are those things best learned?
How do we know they are learned?
The last question is the locus of most of the discussion about the now-omnipresence of generative AI in student work. It is a question of assessment. How do educators know that students are learning the things that are worth learning?
For years, forward-thinking educators have argued that what is worth learning is not memorized facts, but how to think critically, collaborate, and solve real problems. They have argued that assessment should measure understanding, not just compliance. AI has not created this vision. It has made it urgent.
At a dinner discussion this week with a group of educators and education advocates, we explored this idea. The conversation floated between technology, human flourishing, and how fast schools should change in response to rapid advances. We also wondered about the cognitive development of this generation of students. They will never struggle with some of the things we struggled with as young learners. Will they be better off? Or disadvantaged?
One person reflected on writing research papers—something he had done in his formative years with pen and paper, with final drafts typed carefully on a typewriter, even before the days of word processing. He remembered the pride he felt in that effort. The deep thinking that came from developing ideas, gathering evidence, and crafting a coherent argument. All of that can now be replicated by AI in seconds, with no way to detect it.
So, do we go back to pens and typewriters and pretend? Maybe in some spaces. One educator at the table was enthusiastic about a course at his school where all of the writing work happened in paper journals that never left the classroom. That is interesting, and I do think there is value to handwriting and the kind of reflection that comes with it. But for everything? That feels to me like more of a retreat than a solution.
The solutions will come from leaning into the work we have already begun at Keystone. Our teachers are already leading the way, improving instruction and assessment with the tools now available to us. We design assessments that value process over product, thinking over memorization, and integrity over mere originality. We teach students not to avoid AI, but to use it as a tool—while holding fast to the habits of mind that no machine can replicate: curiosity, skepticism, creativity, and care.
AI has not stripped away what is worth learning. It is forcing a reckoning with how we measure it. And our teachers are ready for that reckoning. It is an exciting time for educators and learners alike.
On Monday, Keystone will host the final episode of The Keystone Dialogues, celebrating our 12th anniversary year. I am looking forward to welcoming three distinguished colleagues who have connections to some of the schools in the region that played an important role in Keystone’s development. The Heads of these schools have kept in close touch over the years and even handed down relationships through head of school transitions. It is a meaningful group that has been a great source of support and collegiality to me since my arrival at Keystone, and I am excited to welcome you all into dialogue with them.
Warmly,
Emily
