AI is moving deeper into education, and the next major battleground is not only the student laptop. It is the teacher training room.
OpenAI, Microsoft, and Anthropic have joined a $23 million partnership with one of the largest teachers’ unions in the United States to bring AI into K–12 classrooms. The initiative arrives with big promises, real skepticism, and a central question schools cannot avoid: can AI help students learn without making it easier to avoid learning?
A new push to train teachers on AI
On July 8, OpenAI, Microsoft, and Anthropic announced the National Academy for AI Instruction. The program will train teachers at a New York City headquarters on using AI for classroom instruction and for administrative work such as planning lessons and writing reports.
The training is expected to begin this fall. The effort is led by the American Federation of Teachers, which represents 1.8 million members, and the United Federation of Teachers, which represents 200,000 members in New York.
That scale matters. If the companies and unions succeed in shaping how teachers learn about AI, their influence could extend to millions of educators and students. It also means the curriculum behind the training will carry unusual weight, especially because OpenAI and Anthropic said they could not share specifics despite the launch announcement and the first training being only a few months away.
The companies promoting AI in education argue that it can support more individualized learning, speed up grading, and help teachers create lessons more quickly and creatively. Microsoft representatives have already shown teachers how to use the company’s AI tools for lesson planning and emails, according to the New York Times.
The business case behind the education case
The classroom pitch is not happening in isolation. AI companies are searching for more users, and schools offer a large audience of teachers, students, and institutions that could adopt AI tools as part of daily work.
Anthropic is pitching its AI models to universities, while OpenAI offers free courses for teachers. The National Academy for AI Instruction fits into that broader movement: it presents AI as a professional tool for educators, not just a shortcut used by students.
That distinction is important. Many people currently associate AI in school with cheating, weak attention, and reduced critical thinking. A widely discussed New York magazine story described how constant access to ChatGPT can make it easy to get through college with less effort.
Against that backdrop, AI companies are trying to change the frame. Their argument is that AI can be used responsibly if teachers are trained to understand it, guide it, and integrate it into learning. The challenge is that the companies also have a direct interest in adoption, which makes some educators wary of trusting them to define when AI should be used and when it should be avoided.
What the evidence says so far
The evidence is not one-sided. Some research and surveys do support parts of the case for AI in education.
A recent survey of 1,500 teens conducted by Harvard’s Graduate School of Education found that young people use AI to brainstorm and to ask questions they may not feel comfortable raising in class. Studies in settings including math classes in Nigeria and physics courses at Harvard have also suggested that AI tutors can increase student engagement.
Those findings help explain why schools are interested. A tool that helps students ask more questions, stay engaged, or get more individualized support could be valuable, especially when teachers are already balancing instruction, planning, grading, and reporting.
But the same evidence base also shows the limits of the optimistic case. The Harvard survey found that students often use AI for cheating and shortcuts. An oft-cited paper from Microsoft found that depending on AI can reduce critical thinking. Large language models also produce “hallucinations,” meaning incorrect information can appear as part of their normal operation.
That leaves schools with a difficult balance. AI may help some students learn in some contexts, but it can also make it easier to bypass the work that learning depends on. The source article describes a lack of clear evidence that AI is a net benefit for students, which is why the design of teacher training matters so much.
Educators want literacy, not just prompts
Teachers are already building their own approaches to AI. Christopher Harris, who leads a library system covering 22 rural school districts in New York, has created a curriculum focused on AI literacy.
His lessons cover a wide range of topics. Second graders learn about privacy when using smart speakers. High school students study misinformation and deepfakes.
For Harris, the key is not simply showing teachers a set of tools or pre-written prompts. The goal is a deeper understanding of how AI works, what it can do, and where its limits are. He said the training should produce teachers who understand AI well enough to teach students about the technology itself.
He also pointed to a larger issue that tool training alone cannot solve:
“The bigger issue will be shifting the fundamental approaches to how we assign and assess student work in the face of AI cheating.”
That point goes to the heart of the classroom debate. If assignments and assessments remain unchanged, AI may become a workaround rather than a learning aid. If schools rethink how they evaluate student work, AI could force a broader conversation about what evidence of learning should look like.
Resistance is part of the debate
Not all educators want AI adopted at scale. Several hundred signed an open letter last week resisting its use, and Helen Choi, an associate professor at the University of Southern California who teaches writing, is one of them.
Choi argues that educators should look beyond hype and scrutinize the tools they bring into classrooms. Her position is that large language models should not be widely adopted until they are shown to be useful, safe, and ethical, and until their design reflects educational priorities.
That resistance does not erase the momentum behind AI in schools. But it does clarify the stakes. The question is no longer whether students and teachers will encounter AI. They already are. The question is who defines responsible use, how much evidence schools require before changing classroom practice, and whether teacher training will help educators make independent judgments rather than simply adopt the tools placed in front of them.