Anthropic strengthens school AI safeguards |
Anthropic has agreed to introduce contractually enforceable safety and data privacy protections for school districts using its Claude for Teachers platform, strengthening safeguards around AI in K-12 education. The agreement adopts standards developed through the National Academy for AI Instruction, a five-year, $23m partnership involving the American Federation of Teachers, Anthropic, Microsoft, and OpenAI. Key provisions restrict the use of student and educator data for AI model training, require ongoing breach notification and mitigation, and establish additional oversight for higher-risk technologies, including biometric data collection, profiling, and tracking. School districts must also implement policies and staff training to ensure human oversight of AI-assisted grading, disciplinary decisions, and student placement. While the agreement builds on similar protections adopted by Microsoft, its scope is limited to Claude for Teachers and does not automatically extend to third-party applications using Anthropic’s models. Teachers’ union representatives welcomed the move but emphasized the need for stronger federal regulation, while negotiations with OpenAI remain ongoing.