OpenAI and Anthropic are discussing legally enforceable privacy and safety standards for AI used in schools after Microsoft signed an agreement restricting how student and teacher data can be used.
Microsoft’s agreement with the American Federation of Teachers prohibits school AI systems from using student or educator data for model training, advertising or product development, apart from a narrow safety-related exception.
It also prohibits AI companion features designed to encourage emotional attachment or dependency.
The agreement requires independent audits and plain-language information explaining the protections to students, families and educators.
OpenAI told the Associated Press that it expects to finalise its own agreement with the union.
Anthropic says it is working towards a comparable “gold standard”.
Google has not said whether it will adopt the same framework.
That is significant given the scale of its education products. Google Classroom is used by approximately 150 million teachers and students worldwide.
The negotiations come as schools increasingly adopt generative AI while questions remain over what happens to information students disclose during conversations with the systems.
New York City and Los Angeles have paused student-facing AI deployments while assessing privacy and safety concerns.
Microsoft has now contractually agreed that children’s school-AI conversations will not become training data, advertising material or general product-development data.
OpenAI and Anthropic may soon make comparable commitments.
Sources
- Associated Press – AI companies face pressure over student-data privacy
- Microsoft – National AI Safety & Privacy Standard for schools
Join the Dissenting Citizen Newsletter
Independent news and commentary delivered directly to your inbox.
Breaking stories, analysis and commentary throughout the day.
Follow @VoxDissent →
