AI’s Giants Want to Take Over the Classroom

Author: James O’Donnell (MIT Technology Review) · Published: July 15, 2025 · URL: https://www.technologyreview.com/2025/07/15/1120086/ais-giants-want-to-take-over-the-classroom

One-line: OpenAI, Microsoft and Anthropic put $23 million into training American teachers to use AI — a customer-acquisition move dressed as professional development, launched with no specifics and on evidence that cuts both ways.

Summary

On July 8, 2025, OpenAI, Microsoft and Anthropic announced a $23 million partnership with one of the largest US teachers’ unions to bring AI into K–12 classrooms. The National Academy for AI Instruction would train teachers at a New York City headquarters, starting that fall, both on teaching with AI and on using it for lesson planning and reports.

O’Donnell does not pretend the motive is mysterious. “The companies sponsoring this initiative are, of course, not doing it out of the goodness of their hearts. No — as they hunt for profits, their goal is to make users out of teachers and students.” Anthropic is pitching models to universities; OpenAI offers free teacher courses; at an initial training session, Microsoft representatives showed teachers how to use the company’s AI tools for lesson planning and email.

He also grants they face an uphill battle: the public largely sees classroom AI as “ruinous,” a way to dampen critical thinking, and a viral New York magazine story had described how easy it is to coast through college on ChatGPT.

What the evidence actually says is the piece’s most useful section, and it is genuinely mixed. On the supporting side: a Harvard Graduate School of Education survey of 1,500 teens found kids using AI to brainstorm and to ask questions they’re afraid to ask in class, and studies from maths classes in Nigeria to physics courses at Harvard suggest AI tutors increase engagement. Against: the same Harvard survey shows frequent use for cheating and shortcuts; an oft-cited Microsoft paper found reliance on AI can reduce critical thinking; and hallucination is inherent to how LLMs work. His summary is that “there’s a lack of clear evidence that AI can be a net benefit for students, and it’s hard to trust that the AI companies funding this initiative will give honest advice on when not to use AI in the classroom.”

The specifics problem. Despite the fanfare and a first training months away, “OpenAI and Anthropic told me they couldn’t share any specifics.”

What teachers want instead. Christopher Harris runs a library system covering 22 rural New York school districts and has built an AI-literacy curriculum from smart-speaker privacy for second graders to misinformation and deepfakes for high schoolers. His ask: “The real outcome should be teachers that are confident enough in their understanding of how AI works and how it can be used as a tool that they can teach students about the technology as well” — and the thing to avoid is overfocusing on tools and pre-built prompts teachers use without understanding. But he thinks none of it matters without changing assessment: “The bigger issue will be shifting the fundamental approaches to how we assign and assess student work in the face of AI cheating.”

Scale and resistance. The initiative is led by the American Federation of Teachers (1.8 million members) and the United Federation of Teachers (200,000 in New York); winning them means significant influence over how millions of teachers learn about AI. Several hundred educators signed an open letter against classroom AI. Helen Choi, who teaches writing at USC, is one: “I think it is incumbent upon educators to scrutinize the tools that they use in the classroom to look past hype… Until we know that something is useful, safe, and ethical, we have a duty to resist mass adoption of tools like large language models that are not designed by educators with education in mind.”

Full text

Archived privately against link rot: ../attachments/ais-giants-want-to-take-over-the-classroom/ais-giants-want-to-take-over-the-classroom.md

Connections

• AI Education — the vendor-side push, with the distribution channel (a teachers’ union) named explicitly.
• AI Cheating in Education — Harris’s point that assessment reform is the real dependency, and the Harvard survey finding that the same tool drives both engagement and cheating.
• Students Are Being Treated Like Guinea Pigs Inside an AI-Powered Private School — what the deployment looks like when nobody is holding the vendors to specifics.
• OpenAI and Anthropic — the customer-acquisition logic, stated plainly, and the refusal to share curriculum detail.
• Learning — the underlying question of whether engagement and learning are the same measurement.