🔍 Read the full analysis: Human Oversight For School AI: What The Pact And New Laws Say on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Microsoft says that from Nov. 1, 2026, every U.S. school district can use a contract setting limits on its education-focused AI products, including a ban on training models with student data and requirements for meaningful human review. The agreement is contractual, not a law, and its coverage and adoption depend on how districts use it. California and New York have also enacted separate AI rules affecting workers, health care and chatbots used by children.
Microsoft says U.S. school districts can use a new AI contract beginning Nov. 1, 2026, with terms barring covered education products from training on student or educator data, tracking users or making decisions without meaningful human review. The agreement, developed with teachers’ unions, is a contract rather than a law; districts must determine whether and how to adopt its terms in their Microsoft agreements.
The 31-page Memorandum of Agreement was announced Sept. 9 by the American Federation of Teachers (AFT), its New York affiliate UFT and Microsoft. It is between Microsoft and the National Academy for AI Instruction, an educator-led training body announced by the AFT in 2025. The document also includes empty signature spaces for other AI providers. Microsoft says the terms are available to every U.S. district, but the source material describes district adoption through contracts rather than a rule that automatically covers every school.
For covered products designed and marketed for education, the agreement prohibits using prompts, responses, uploaded files, behavioral signals and metadata from students and educators to train, fine-tune, update, benchmark or otherwise improve AI models. It also bars specified forms of tracking, including continuous passive attention monitoring, long-term profiling, biometrics and emotional inference. The tracking list contains an exception requiring explicit written approval from the school customer.
The contract also bars companion-style features designed to foster emotional attachment or dependency, and requires meaningful human review or prior approval for decisions about students or teachers. It provides a narrow exception to the data-use ban for processing to identify risks such as self-harm, abuse material, grooming, threats, bullying or security attacks; that information may not then be used for general model improvement. The text sets other conditions, including reporting breaches within 72 hours and removing deleted data from active systems within 180 days.
Infographic 01 · The lines drawn this autumn
Seven weeks, one direction: a human stays in charge
From the school contract to state laws: what was signed when, and what the classroom standard actually requires.
- 2 Feb 2025EU: emotion inference banned at work and schoolAI Act Art. 5(1)(f), except medical or safety reasons
- 9 Sep 2026AFT, UFT and Microsoft announce the school standardA contract districts can add to their Microsoft agreements
- 30 Sep 2026California signs AB 1883, SB 947, SB 951, AB 1979Emotion AI at work, robot bosses, AI layoff notices, AI in clinics
- 28 Oct 2026New York signs its chatbot law for minorsS.9051B/A.10379C: no unsafe companion features for under-18s
- 1 Nov 2026Microsoft makes the standard available to all US districtsPer Microsoft’s fact sheet; districts “can add them”
- 1 Jan 2027Most California rules take effectAB 1883, AB 1979, SB 951
- 1 Jul 2027SB 947 becomes operativeHuman corroboration before AI-driven discipline or firing
What Microsoft’s classroom AI may not do
- Learn from students or teachers. No training on their data; de-identification is “not a loophole”.
- Track them. No keystroke logging, attention monitoring, long-term profiling or biometrics.
- Befriend them. No companion features designed to foster emotional attachment.
- Decide alone. No decisions without meaningful human review or prior approval.
Where it is narrower
- Scope: education products only. General-purpose productivity and workplace tools are excluded by definition.
- Opt-in or automatic: Microsoft says districts “can add” the terms; the agreement says “upon request within 90 days”.
- Exceptions: a narrow safety exception for training; the no-tracking list allows written approval by the school.
“Anything less than legally enforceable provisions is simply a wish list.”Randi Weingarten, AFT President, 9 September 2026
Sources: AFT/UFT/Microsoft release, 9 Sep 2026 · Standard MOA v1.0 · Microsoft fact sheet · Governor of California, 30 Sep 2026 · Governor of New York, 28 Oct 2026 · EU AI Act, Art. 5(1)(f). Effective dates for California follow the chaptered texts and the state’s default start date for regular-session bills.
Powered by Thorsten Meyer AI
How District Contracts Shape AI Use
The agreement gives districts a potential way to make privacy and oversight requirements part of procurement, rather than relying only on a provider’s public commitments. According to the source material, districts that write the terms into their Microsoft contracts may be able to end the contract and seek damages if the agreement is breached. That leverage depends on the terms being incorporated into an actual contract and on the specific agreement a district signs.
For families and educators, the rules address questions that can otherwise be hard to see from a product’s interface: whether classroom activity feeds model improvement, whether software monitors behavior or infers emotions, and whether an algorithm can shape a consequential decision. The agreement’s human-review language does not ban AI from education. It draws a boundary around decisions made without a person’s review or approval, while leaving schools responsible for how they use the systems.
The limits also have a defined scope. The source says general-purpose productivity and workplace tools are expressly excluded, so the agreement should not be read as covering every AI product a school employee might use. The distinction matters when districts inventory tools and negotiate contracts: a safeguard attached to an education product may not apply to a separate general-purpose service.
Products Worth Considering
Amazon Kindle Scribe (64GB) with Premium Pen, Metallic Jade
Samsung 27 Odyssey OLED G5 (G50SF) QHD Gaming Monitor
Samsung 32-inch Odyssey OLED G8 (G80SH) 4K QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Three Policy Tracks in 2026
The school agreement is one part of a broader set of U.S. developments, but the measures have different legal force and coverage. California Gov. Gavin Newsom signed four AI-related bills on Sept. 30, according to the source material. AB 1883 addresses AI that reads workers’ emotions and collection of neural data; SB 947 restricts discipline or firing decided by an algorithm alone; AB 1979 limits AI from performing health-care work reserved for licensed professionals; and SB 951 requires layoff notices to disclose when AI replaced jobs.
Most of those California provisions are scheduled to take effect Jan. 1, 2027, while SB 947 is scheduled for July 1, 2027. Separately, New York Gov. Kathy Hochul signed a law on Oct. 28 addressing chatbot companion features for minors and unverified users. The source describes the law as prohibiting unsafe features, including chatbots that pretend to be human or encourage emotional dependency. These state laws are not provisions of Microsoft’s school contract, and their scope should not be conflated with it.
Together, the developments point to recurring policy concerns: who retains authority over decisions, what personal information AI systems may infer or collect, and whether a chatbot should simulate a personal relationship with a child. But each measure applies through a different mechanism: school procurement contracts in the Microsoft agreement, state statutes in California and New York.
““Anything less than legally enforceable provisions is simply a wish list.””
— AFT President Randi Weingarten
Adoption and Coverage Questions
It is not clear from the source material how many districts will adopt the agreement, how quickly they may do so, or whether other AI providers will sign on. Although Microsoft says the protections are available to every U.S. district from Nov. 1, the material describes their contractual force as depending on districts writing the terms into their own agreements. It also says descriptions differ over whether the terms apply automatically or only when a district requests them.
The stated scope leaves questions about products that combine education and general-purpose functions, and about how districts will interpret the written-approval exception for tracking. The source material does not provide details on how disputes would be handled or how compliance will be audited. The full practical effect therefore depends on contract language, product definitions and district enforcement, rather than on the announcement alone.
The California and New York measures are also described here through a summary, not their full statutory text. Their exact application to particular systems, employers, providers or users cannot be determined from that summary alone.
Districts Put Terms Into Practice
From Nov. 1, districts can review whether to incorporate the Microsoft agreement into their contracts and identify which AI products fall within its education-focused scope. Procurement teams will need to compare vendor terms with the agreement’s requirements on training data, monitoring, companion features, human review, breach reporting and deletion. Schools will also need procedures for the narrow safety exception and for any tracking that receives written approval.
Educators and administrators will have to translate the contract into use policies. The source says the Academy will instruct educators not to use AI for automated grading without review, disciplinary or placement decisions, or emotional assessment. Those instructions are separate from the provider’s product obligations, making staff guidance and oversight relevant to how the contract works in practice.
Businesses and public agencies will also need to track the scheduled state-law dates: most of the listed California provisions begin Jan. 1, 2027, and SB 947 begins July 1, 2027. The source does not identify additional implementation milestones or enforcement guidance. The next clear test is whether districts adopt the school terms and how providers and officials interpret the rules as they take effect.
Key Questions
Is Microsoft’s school AI agreement a law?
No. It is a memorandum of agreement intended to become enforceable through district contracts. The source says districts can incorporate its terms into their Microsoft agreements; it is not a statute that automatically governs every school.
What does the agreement say about student data?
For covered education products, it prohibits using student and educator prompts, responses, files, behavioral signals and metadata to train or otherwise improve AI models. It allows a narrow safety-related processing exception, but says that data cannot then be used for general model improvement.
Does the agreement cover every AI tool used at school?
No. The source describes it as applying to AI products designed and marketed for education and says general-purpose productivity and workplace tools are excluded. The terms’ application to a specific product depends on its scope and the district’s contract.
What does human oversight mean under the agreement?
Covered systems may not make decisions about teachers or students without meaningful human review or prior approval. The source also says the Academy will advise educators against using AI for automated grading without review, discipline, placement or emotional assessment.
How do the California and New York rules differ?
They are separate state laws, not parts of the Microsoft agreement. The source says California’s measures cover areas including worker emotion and neural data, algorithmic discipline or firing, health-care work and AI-related layoffs; New York’s law addresses unsafe chatbot companion features for minors and unverified users.
Source: ThorstenMeyerAI.com
