The tech industry’s campaign to embed artificial intelligence chatbots in classrooms is accelerating.

The American Federation of Teachers, the second-largest U.S. teachers union, said Tuesday that it would start an AI training hub for educators with $23 million in funding from three leading chatbot makers: Microsoft, OpenAI and Anthropic.

The union said it planned to open the National Academy for AI Instruction in New York City, starting with hands-on workshops for teachers this fall on how to use AI tools for tasks like generating lesson plans.

Randi Weingarten, president of the American Federation of Teachers, said the AI academy was inspired by other unions, such as the United Brotherhood of Carpenters, that have worked with industry partners to set up high-tech training centers.

The New York hub will be “an innovative new training space where school staff and teachers will learn not just about how AI works, but how to use it wisely, safely and ethically,” Weingarten said in an interview. “It will be a place where tech developers and educators can talk with each other, not past each other.”

The industry funding is part of a drive by U.S. tech companies to reshape education with generative AI chatbots. These tools, including OpenAI’s ChatGPT and Microsoft’s Copilot, can produce humanlike essays, research summaries and class quizzes.

In February, California State University, the largest U.S. university system, said it would provide ChatGPT for some 460,000 students. This spring, Miami-Dade County Public Schools, the third-largest U.S. school district, began rolling out Google’s Gemini AI for more than 100,000 high schoolers.

The Trump administration, which recently froze nearly $7 billion in funding for schools, has called on industry to pony up for AI education. Last week, the White House urged U.S. companies and nonprofit groups to provide AI grants, technology and training materials for schools, teachers and students. Since then, dozens of companies have signed on, including Amazon, Apple, Google, Meta, Microsoft, Nvidia and OpenAI.

Some tech executives hope AI will become the fourth R.“Reading and writing and arithmetic and learning how to use AI,” said Chris Lehane, OpenAI’s chief global affairs officer. “You’re going to have to learn those skills over time, and I do think our education system is the best place to be able to do that.”

But some researchers have warned that generative AI tools are so new in schools that there is little evidence of concrete educational benefit — and significant concern about risk.

Chatbots can produce plausible-sounding misinformation, which could mislead students. A recent study by law school professors found that three popular AI tools made “significant” errors summarizing a law casebook and posed an “unacceptable risk of harm” to learning.

Outsourcing tasks including research and writing to AI chatbots may also hinder critical thinking, a recent study from Microsoft and Carnegie Mellon University found.

“I do think that there is a risk,” said Brad Smith, president of Microsoft, noting that he frequently cited the critical thinking study to employees. He added that more rigorous academic research on the effects of generative AI was needed. “The lesson of social media is don’t dismiss problems or concerns.”

Union experts have also raised alarms about the industry’s practices. Some tech companies have trained their AI models on swaths of texts scraped from the internet, without compensating writers and other creators, or outsourced the labeling of training data to low-paid workers.

(The New York Times has sued OpenAI and its partner, Microsoft, over copyright infringement of news content. Both companies have denied wrongdoing.)

Trevor Griffey, a lecturer in labor studies at UCLA, also warned that tech firms could use AI deals with schools and the teachers union as marketing opportunities to make students lifetime chatbot customers.

“It’s a long-game investment by companies to turn young people into consumers who identify with a particular brand,” said Griffey, a vice president of University Council-AFT Local 1474, a union representing University of California librarians and lecturers.

Weingarten said she was aware of the concerns and that her union, which represents 1.8 million members, had developed AI school use guidelines to address some of them.

The union’s new training hub will be in the downtown Manhattan headquarters of the United Federation of Teachers, which represents nearly 200,000 New York City teachers and other school employees.

Microsoft will provide $12.5 million for the training effort over the next five years, and OpenAI will contribute $8 million in funding and $2 million in technical resources. Anthropic will add $500,000 for the first year of the effort.