# Teachers got AI guardrails in writing, not promises

> Two teaching unions skipped Washington and signed a legally enforceable AI deal with Microsoft instead.
- Title: Teachers got AI guardrails in writing, not promises | Shortify
- Summary: Two teaching unions skipped Washington and signed a legally enforceable AI deal with Microsoft instead. When federal rules stall, the people closest to the…
- Keywords: ai in education, privacy, microsoft, teachers unions, policy, technology, Teachers, got, guardrails, writing, promises, TechRadar
- Source: TechRadar — https://techradar.com/pro/microsoft-is-working-with-the-american-federation-of-teachers-to-work-out-how-best-to-use-ai-in-the-classroom
- Published: 2026-09-13T16:20:00+00:00
- Read time: 1 min
- Topics: ai in education, privacy, microsoft, teachers unions, policy, technology
## Washington stalled, unions moved
Federal legislation to limit AI overreach in schools failed. So the American Federation of Teachers and the United Federation of Teachers cut their own deal with Microsoft.
## A standard with teeth
The National AI Safety & Privacy Standard is legally enforceable, not a set of principles. It came out of months of negotiation between the unions and Microsoft.

> anything less than legally enforceable provisions is simply a wish list
## Student data stays out
Microsoft agreed not to use student or teacher data to train or improve AI models. The data also cannot be sold, monetized, or repurposed.
## Schools hold the controls
Districts decide which AI tools run in classrooms, and how data is stored and deleted. Providers must flag product changes and new risks as they happen.
## Parents get to see inside
The agreement requires transparency about what data is collected and what protections apply, so families are not guessing about what a classroom tool does.

> protect the privacy and data of children, give schools real control over AI tools in the classroom, and provide transparency and information to parents
## Humans stay in the loop
AI systems covered by the standard need human oversight and must be built to block harmful actions and shield sensitive information.
## Cities already pulled back
New York City put a moratorium on AI in its schools and Los Angeles Unified did much the same. NYC has now adopted the union framework banning screens and AI in the younger years.
## OpenAI and Anthropic are next
The AFT says it is in talks with both companies to sign the same standard. Brad Smith says Microsoft will extend it to every district in the country.

> we'll extend this agreement to every school district across the country
## Key takeaway

When federal rules stall, the people closest to the problem can write binding ones themselves, and a contract beats a policy statement.