The Contract Was Signed. The Conversation Wasn't.

This week NPR walked through California State University's renewed deal with OpenAI. The Chancellor's Office signed a contract to give every one of its more than 470,000 students and 63,000 faculty and staff access to ChatGPT Edu, a commercial product purchased for $13 million a year over three years, the largest agreement OpenAI has signed with any university in the country.¹

The number is the headline. The sentence that lingers is from Katie Karroum, vice president of systemwide affairs for the Cal State Student Association: "We were not consulted when the contract was signed, and we weren't even given a heads up."²

That detail is doing a lot of quiet work.

What the Survey Asked

A systemwide CSU survey of more than 94,000 students and university employees, released earlier this spring, came back with an unsettling pair of numbers. Fifty-two percent of faculty said AI was having a negative effect on their teaching. Sixty-seven percent of students said their professors don't teach them how to use AI effectively.³

The juxtaposition is what matters. The same people being given the most expensive AI subscription in higher education are saying the institution has not done the teaching that would make the subscription meaningful. The tool arrived ahead of the pedagogy. The license arrived ahead of the curriculum. The contract arrived ahead of the conversation.

Chancellor Mildred García called the survey results "a call to action."³ Faculty read the same results and delivered a petition titled "Cancel ChatGPT Edu. Invest in Humans." It has gathered more than 3,300 signatures.⁴ Martha Kenney, a professor at San Francisco State, put the refusal directly to NPR: "I think refusing this technology needs to be a position that's on the table."¹

What Rollouts Skip

The pattern is familiar to anyone who has lived through an institutional technology rollout. The decision is made at the top, often for defensible reasons: scale, equity of access, a worry about being left behind. The deployment follows. Faculty learn about the new tool when their inbox tells them they have an account. Students discover it through a banner on the LMS. Somewhere later, after the tool is already part of the building, the question of how it should be used in teaching shows up as a workshop on a Thursday afternoon.

What is skipped is the part that matters most. The slow, specific, classroom-level work of figuring out what changes for the people in the room. What changes about how a student drafts a paper. What changes about how a teacher gives feedback. What changes about what a course is for.

Refusing technology, as Kenney suggests, is one legitimate response to that gap. It is not the only one. Another is to insist that any tool brought into a classroom be brought in with the people who will be in the classroom. Not consulted in the legal sense. Consulted in the working sense. With time. With drafts. With the willingness to redesign assignments instead of layering a new tool on top of old ones.

What the Future Could Look Like

The future worth imagining is not a future without these tools. Students are already using them. Ninety-five percent of CSU students said they had used an AI tool of some kind.³ The question is whether the tools live inside the work of teaching or beside it.

Inside the work means a few things. It means the moments when a student turns to an AI tutor are visible to the teacher, not as surveillance, but as evidence of where the student got stuck. It means a draft remembers itself, that the version a student wrote at the kitchen table at nine on Sunday and the version she pasted in at noon on Monday are both still there, and the difference between them is the thing the teacher learns from. It means process stops being a checkbox at the end of a project and starts being the actual material the class reflects on.

This is the part a system like Koan tries to be useful for, in a quiet way. It holds onto the drafts a student kept, the questions she asked Aidan when she was stuck, the pauses that mean something and the pauses that do not. None of that, in itself, is teaching. It is closer to the raw evidence teaching has always needed and rarely had time to collect.

The Sentence Underneath

The reason the line about consultation matters is that it tells you what kind of authority the new tools are being given. If the largest AI rollout in higher education can begin without asking the people inside its classrooms, the template for every smaller rollout has been set. The template is: buy first, teach later, hope it lands.

There is another template. Ask first. Teach with. Decide together what the tool is for. It is slower. It is also the only template that has ever turned a piece of technology into a piece of education.

If a technology arrives in your classroom faster than the conversation about why it is there, who is it actually serving?