Earlier this month, OpenAI, Microsoft, and Anthropic announced a $23 million partnership with major U.S. teachers’ unions to bring AI into public K–12 education. The National Academy for AI Instruction, as it’s called, will train teachers starting this fall in New York City, helping them use AI for lesson planning, grading, and communication.

This marks one of the largest investments to date in AI-specific teacher training. But it also raises familiar questions — about trust, effectiveness, and who ultimately benefits.

Let’s break it down.

🧭 The Vision — and the Tensions

The initiative is backed by the American Federation of Teachers (1.8 million members) and the United Federation of Teachers (200,000 members). Its stated goal is to empower educators with AI, not replace them.

Supporters say AI can:

  • speed up planning and admin tasks,
  • support more creative or differentiated instruction,
  • help students brainstorm ideas or ask hard questions more freely.

But critics — including hundreds of educators who recently signed an open letter — warn of potential downsides:

  • overdependence on tools designed by tech companies, not teachers,
  • reduced critical thinking,
  • the growing challenge of AI-fueled cheating,
  • hallucinated or misleading content.

Even the Harvard Graduate School of Education’s own data reflects this tension: while many teens use AI to explore ideas, just as many use it to avoid doing the work.

🧑‍🏫 The Role of Teachers Can’t Be Optional

As Christopher Harris, a school library director involved in AI curriculum development, noted:

“The real outcome should be teachers confident enough in their understanding of how AI works […] so they can teach students about the technology as well.”

We fully agree.

Teaching with AI is not the same as teaching about AI — and both are essential.

The risk of these large-scale initiatives is that they may emphasize tools over understanding, productivity over pedagogy.

👩‍🏫👨‍🎓 AI as a Shared Space: Teacher–Student Collaboration First

At GoSchool, we believe that generative AI represents an unprecedented opportunity in human history:

for the first time, learners have direct access to the written (and increasingly multimodal) knowledge accumulated by humanity.

We believe this democratization of knowledge can enable more equitable, engaging, and effective learning than ever before.

But only if teachers and students use AI together — not separately.

That’s why our platform is built around:

  • teacher-curated course content as the only knowledge source for AI,
  • teacher supervision and human oversight for every AI answer,
  • and classroom workflows that keep student agency and collaboration front and center.

🎓 Learning Happens in Relationships, Not Interfaces

We don’t need to choose between AI optimism and AI fear.

We need to rethink learning as a collaboration between teacher, student, and AI — with pedagogy, trust, and transparency leading the way.

If the National Academy for AI Instruction embraces this mindset, it has a real chance to help—not just automate—education.

Because at GoSchool, we believe the classroom doesn’t end with the bell.

With the right guidance, AI can extend the classroom — and make learning more human than ever.

Learn more about GoSchool’s approach to teacher–student AI collaboration at goschool.ai or contact us to bring our platform to your institution.

Source

This post responds to reporting from MIT Technology Review:

🔗 AI’s Giants Want to Take Over the Classroom – Technology Review (2025.07.15)