Abstract:
Microsoft announced new school artificial intelligence security and privacy standards with the American Federation of Teachers (AFT) and its New York City affiliate, the New York City Union of Teachers (UFT). Under the agreement, school districts that adopt the standards can put provisions into their contracts that will make them legally binding, including not using student or educator data to train AI models, minimizing data collection, providing plain language to families about how the tools work, prohibiting the deployment of AI companion products, and implementing human review of "high-risk" decisions.
Not long before this agreement was introduced, two major public school districts in New York City and Los Angeles announced that they would temporarily ban AI tools for many students for one year in order to evaluate the reasonable use of such tools on campus and the necessary safety protection measures. As concerns among parents and some educators grow about the introduction of AI in schools and increased screen time for students, Microsoft is apparently hoping to reduce the possibility of further tightening of policies by school districts through new binding rules.
Starting this November, school districts will have the option to add these terms to new or existing Microsoft partnership agreements without having to renegotiate the entire contract. One of the core requirements of the agreement is that Microsoft must not use student and teacher data for model training; at the same time, the company must limit the amount of raw data collected and clearly explain to student families how its AI products process and use relevant information.
The American Federation of Teachers has long been cautious about the introduction of educational technology into schools, but it has not completely rejected cooperation with technology companies to help educators cope with the rapid development of AI. Randy Weingarten, the organization's president, argued in May that screen devices should be banned for students under third grade and that AI tools for students should not be used until middle school. However, AFT also previously announced the establishment of an AI training center for educators, receiving a total of $23 million in funding from Anthropic, Microsoft, and OpenAI.
When talking about the agreement with Microsoft, Weingarten said the terms were "rock solid" and necessary because "no other agency, including the federal government, stepped up to do the work that really needed to be done." She believes that in the face of risks caused by campus AI, the education sector cannot just express dissatisfaction, but should take decisive action; any commitment lacking legal enforceability is just a "wish list."
Comments