When we published Taking Student AI Resistance Seriously, we laid out 22 concerns students raise about AI, including cheating, water use, and military contracts. For each one, we said whether better practice can solve it, whether a fix is still emerging, or whether the concern stands. We have now added a 23rd concern, that AI could cause human extinction, under a new topic called Existential.
Why we added it
In early September, a researcher who had worked on pretraining at both OpenAI and Anthropic resigned and wrote publicly that the labs are "gambling with our lives." The next day, Anthropic's head of alignment science agreed in public and put his own estimate of catastrophe at more than ten percent within a decade. Within a week, the chief executives of Anthropic and OpenAI were both calling to "pace the frontier," and Microsoft AI released a draft code of conduct stating that "AI should not exceed human control."
These warnings came from the people building the systems, and their companies have kept building them.
What the entry says
A student who reads those statements and decides to keep their distance from AI is taking the builders at their word. We mark the concern as one that stands, because better classroom practice does not resolve it.
The entry does not estimate the odds, because the people closest to the work disagree sharply about them. It also makes no claim that one student's choice to avoid a chatbot will change what a frontier lab trains next.
Where students and campuses can act
Students and campuses still have some influence here. The industry has named its own proposed safeguards in public, and several of them pass through places a campus can reach.
The first step the labs proposed is independent evaluation by outside experts with deep access to frontier systems, and those evaluators come from universities. Microsoft released its draft code for public comment, so a student, a department, or a campus AI committee can respond to codes like it. Laws on evaluation access, disclosure, and compute reporting would make a promise to slow down enforceable.
The resignation that started this was one employee's decision, and some of today's students will be the employees inside these companies.
For campus leaders
If your campus has encouraged students to adopt AI, the students who share this concern deserve an answer when they ask why they should use it, and that answer needs more than a usage policy. Concern 23 and the other 22 entries are in the full field note.
Tagged