🔍 Read the full analysis: What Do Microsoft’s School AI Pact And California’s New Laws Mean For Human Oversight? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Microsoft says that from Nov. 1, 2026, U.S. school districts can use a contract barring covered education AI products from training on student data, tracking students or making decisions without human review. The agreement is contractual, not a law, and its scope and adoption remain important limits; California and New York have separately enacted laws covering some workplace, health care and child-facing AI uses.
Microsoft says it is making a school AI contract available to every U.S. school district from Nov. 1, 2026, setting limits on how covered education products can use student information and make decisions. The contract is part of a wider run of policy changes: California has signed laws regulating certain workplace and health care uses of AI, while New York has enacted rules for some chatbots used by children. The measures do not prohibit AI broadly; they set boundaries for particular products and uses, with different legal force and scope.
The school agreement is a 31-page memorandum of agreement between Microsoft and the National Academy for AI Instruction, an educator-led training body announced by the American Federation of Teachers in 2025. The source says the document also has empty signature slots for other AI providers. Districts can incorporate its terms into their Microsoft contracts, giving them a contractual route to seek remedies if the terms are breached. It is not a federal law or a general Microsoft policy applying to every product.
For covered education AI products, the agreement bars using student, teacher and administrator data to train or improve AI models, with a narrow exception for processing data to identify specified safety risks. It also restricts location and behavioral tracking, keystroke logging, passive attention monitoring and emotional inference. Providers may not offer companion-style features designed to foster emotional attachment, and covered systems may not make decisions about students or educators without meaningful human review or prior approval. The document provides for breach reports within 72 hours and deletion from active systems within 180 days.
The source describes four California laws signed by Gov. Gavin Newsom on Sept. 30. AB 1883 addresses AI emotion reading and neural data at work; SB 947 bars discipline or firing decisions made by an algorithm alone; AB 1979 limits AI from performing tasks reserved to licensed health professionals; and SB 951 requires layoff notices to disclose when AI replaced jobs. Most are scheduled to take effect Jan. 1, 2027, while SB 947 is scheduled for July 1, 2027. New York Gov. Kathy Hochul signed a law on Oct. 28 addressing what the source calls unsafe companion features for minors and unverified users.
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
Where Human Review Is Required
The measures matter because they focus on specific points where automated systems can affect people: handling student information, monitoring behavior, deciding discipline or employment outcomes, and interacting with children as though they were human companions. For schools, the Microsoft agreement could provide procurement teams with contract language to demand privacy and oversight protections from providers. Its practical reach, however, depends on districts adopting the terms and on which products fall within the definition of education AI.
For employers and health care providers, California’s laws create a separate compliance task. Organizations will need to determine whether their tools infer emotion or collect neural data, whether software plays a role in employment decisions, and whether an AI product is performing work reserved for a licensed professional. Those are not blanket restrictions on workplace or clinical AI; the source describes targeted limits, with different effective dates.
Together, the developments illustrate a policy approach based on human accountability rather than a general ban on AI. That distinction is relevant to workers, students, teachers and parents: protections may exist on paper, but their availability will depend on the law’s coverage, a district’s contract choices and how a particular system is used.
Products Worth Considering
Bose QuietComfort True Wireless Noise-Cancelling Earbuds, Black
Apple MacBook Air 15-inch (2026) with M5 chip, 16GB memory, 512GB SSD, Silver
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
As an affiliate, we earn on qualifying purchases.
Three Different Routes to Oversight
The Microsoft agreement was announced on Sept. 9 by the AFT and United Federation of Teachers with Microsoft. Microsoft says districts can access it beginning Nov. 1. The source attributes two defenses of the agreement to its supporters: AFT President Randi Weingarten said, “anything less than legally enforceable provisions is simply a wish list,” while Microsoft President Brad Smith described it as a standard that “sets a high bar for child privacy and AI safety.” These are their characterizations, not independent findings about how the terms will work in practice.
The agreement’s enforcement route differs from legislation: a district can put its provisions into a Microsoft contract and, under the source’s description, may end the contract or seek damages if terms are broken. The document is limited to AI products designed and marketed for education. It expressly excludes general-purpose productivity and workplace tools, so it should not be read as a universal ban on using AI in schools or as a rule governing every tool a student might encounter.
California’s provisions operate through state law, with staggered start dates, while New York’s law addresses certain chatbot features for children. The source groups these actions around human review, limits on what AI can infer about people, and restrictions on systems designed to simulate emotional relationships. The common theme does not make their legal requirements interchangeable.
““anything less than legally enforceable provisions is simply a wish list.””
— Randi Weingarten, president of the American Federation of Teachers
Coverage and Enforcement Questions
How widely districts will adopt the agreement is not stated, nor does the source establish whether Microsoft applies the terms automatically or only when a district requests or negotiates them. It notes that Microsoft, the agreement and press reports describe this point differently. The document’s contractual route also means its protections should not be presented as statutory rights that automatically apply in every U.S. school.
The source says the agreement allows written school approval for exceptions to some tracking restrictions, but does not explain how frequently districts might grant them or how disputes over compliance would be handled. Details about independent auditing and enforcement outcomes are not provided. California’s laws and New York’s chatbot law are also described without full statutory text here, so specific definitions, exceptions and enforcement procedures cannot be assessed from the supplied material alone.
It remains unclear how the provisions will apply to particular products and mixed-purpose tools, especially where an education product includes general-purpose features. The scope of the New York law’s terms, including which chatbot features qualify as unsafe and how “unverified users” are treated, also requires reference to the law itself and subsequent implementation guidance.
District Contracts and 2027 Deadlines
Starting Nov. 1, districts can consider whether to include the Microsoft agreement in their contracts and how its restrictions fit their existing technology purchases. Providers may also face questions from district procurement teams about data use, tracking, companion features and human review. The source says the agreement leaves room for other AI providers to sign on, but it does not report that any have done so.
California’s listed effective dates are the next fixed milestones: most of the four laws take effect Jan. 1, 2027, with SB 947 following on July 1, 2027. Employers, health care organizations and technology suppliers will need to assess the final requirements and their own uses before those dates. The next practical indicators will be how state agencies interpret the laws and whether organizations change procurement, monitoring or decision-making practices.
For New York, the law was signed Oct. 28, but the supplied material does not give an effective date or implementation schedule. Further guidance and details about enforcement will clarify how the restrictions apply to child-facing chatbots. Until then, the confirmed development is a set of targeted rules and contract terms—not a single national standard covering every use of AI.
Key Questions
Is Microsoft’s school AI agreement a law?
No. It is a memorandum of agreement that districts can put into Microsoft contracts. Its enforcement depends on contractual terms, unlike California’s state laws.
What does the agreement prohibit?
For covered education AI products, it restricts model training on student and educator data, specified forms of tracking and emotional inference, companion-style features, and decisions without meaningful human review. The source describes a narrow safety-processing exception for specified risks.
Does the agreement cover every Microsoft AI product used in schools?
No. The source says it covers products designed and marketed for education and excludes general-purpose productivity and workplace tools. How particular mixed-purpose products will be treated is not specified.
When do California’s AI laws take effect?
According to the source, most of the four laws signed Sept. 30 take effect Jan. 1, 2027. SB 947, concerning algorithm-only discipline or firing decisions, is scheduled to take effect July 1, 2027.
What do the New York rules cover?
The source says the law signed Oct. 28 addresses certain unsafe companion features in chatbots used by minors and unverified users. It does not provide the law’s effective date or full implementation details.
Source: ThorstenMeyerAI.com
