Teachers Unions and Microsoft Seal First Nationwide AI Safety and Privacy Standard for U.S. Schools

The binding Memorandum of Agreement, signed by Weingarten and Smith on September 7, 2026, lets school districts embed ten enforceable AI-privacy and safety principles into their Microsoft contracts. The standard bars Microsoft and any signatory provider from using “Covered Data”—prompts, outputs, uploaded files, behavioral signals and metadata—to train or improve AI models, except for narrowly defined safety and security purposes. Districts may seek damages and pause deployments if a provider breaches the rules.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools Mother JonesFortune MagazineABC7 New York

The agreement builds on AFT-led policy work, including the 2025 launch of the National Academy for AI Instruction. President Weingarten’s May 2024 speech called for a screen ban in kindergarten through second grade and a ban on student-facing AI in elementary schools. Similar frameworks have been adopted by Los Angeles Unified, New York State United Teachers, New York City Public Schools, and Mayor Zohran Mamdani’s administration.

ABC7 New York

• September 7, 2026 – Memorandum of Agreement signed.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• September 9, 2026 – Public announcement of the National AI Safety & Privacy Standard.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• November 1, 2026 – Microsoft states the protections will be available to every U.S. school district.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• December 31, 2027 – Addendum targets ISO 42001 certification for Microsoft’s education AI products.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• Ten binding principles covering data use, collection limits, ownership, human oversight, security certifications, breach reporting, and export rights.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• 180-day deletion deadline for any student or teacher data on Microsoft’s active systems after a district request.

Mother Jones

• 72-hour breach-notification requirement to districts.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• Annual independent penetration testing and certifications (SOC 2 Type II, ISO 27001, ISO 27701, ISO 42001 or equivalent, FedRAMP Moderate where required).

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• Data residency limited to the EU or North America; de-identified telemetry excluded from model-training bans.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• Two-year term with automatic renewal only upon written agreement; providers may withdraw with 60 days’ notice.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

Mayor Mamdani reiterated the city’s one-year moratorium on student-facing AI for younger learners and framed the policy as prioritizing students over profit.

Mother Jones

• Effective date: Fortune reports the protections took effect for all districts on November 1, while Unite’s fact sheet lists November 1, 2026 as a scheduled rollout.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• Tracking exception: The addendum permits limited data use for Microsoft’s “Speaker Coach” and “Speaker Progress” tools, a detail not mentioned elsewhere, leaving the scope of that exception unclear.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools

• “I want to have safeguards in law to ensure that AI is used for its promise and that we guard against its dangers. This is a very important first step,” — Randi Weingarten, AFT president

Fortune Magazine

• “It is 30 pages, but in some ways it comes down to three words: privacy, safety, and transparency,” — Brad Smith, chair and president

Fortune Magazine

Microsoft plans to make the protections available to every U.S. school district beginning November 1, 2026, regardless of formal adoption. The addendum requires ISO 42001 certification by December 31, 2027. The AFT continues discussions with OpenAI and Anthropic about joining the framework, and districts that sign on will receive plain-language guides for families within 90 days of the effective date.

Microsoft Signs Binding AI Safety and Privacy Standard for US Schools