Privacy
Privacy Watch: Claude Ban, School Moratorium, Wrongful Arrest, Apple CSAM Suits
Microsoft bans Claude over data retention; NYC council pushes AI pause; Florida man sues after AI arrest; Apple hit with $1.2B CSAM lawsuit.
Microsoft Bans Claude Fable 5 Over Anthropic Data Retention Rules
Microsoft has temporarily banned employees from using Anthropic’s Claude Fable 5 AI model following new data retention policies. Anthropic can now retain prompts and outputs for up to 30 days, with flagged content held for up to two years, raising concerns that sensitive corporate and customer data could be exposed to a competitor. Microsoft’s legal and compliance teams are reviewing the changes, while the company also plans to terminate all Claude Code licenses by June 30, pushing employees toward its own GitHub Copilot CLI. The timing coincides with Microsoft’s fiscal year-end, suggesting a cost-cutting motivation alongside data security worries.
NYC Council Members Demand Two-Year AI Moratorium in Schools
Twenty-nine New York City council members sent a letter to Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels on June 9, 2026, calling for a two-year pause on artificial intelligence in the nation’s largest school system. The letter criticizes the NYC Department of Education’s AI guidance as flawed for failing to strengthen student data privacy protections. Nationally, a coalition led by Fairplay has pushed for a five-year halt on student-facing generative AI products, citing privacy risks. The council members also pointed to a state comptroller audit showing that NYCPS policies do not fully align with cybersecurity standards, and warned of broader harms to cognitive development and mental health. The American Federation of Teachers has separately called for a ban on student-facing AI tools in elementary schools.
NYC schools face public pressure to pause AI use for 2 years →
Florida Man Sues After AI Facial Recognition Leads to Wrongful Arrest
Robert Dillon, a 52-year-old Fort Myers resident, has filed a lawsuit with the ACLU after being arrested and prosecuted for attempted child luring based on a faulty AI facial recognition match. Jacksonville Beach police used Faces software that flagged Dillon as a 93% match for a suspect captured on low-quality security footage at a McDonald’s, despite Dillon living over 300 miles away and never visiting the area. Prosecutors dropped all charges in August 2024, but Dillon says the ordeal destroyed his reputation and left him afraid to interact with children. The lawsuit accuses lead investigator Scott O’Connell of ignoring exculpatory evidence, including automated license plate data proving Dillon’s vehicles were never near the scene. Legal documents note that Dillon is at least the 15th person nationwide to be wrongly arrested due to faulty AI matching, highlighting the lack of regulation in police use of facial recognition.
📡 From social media
Apple Faces $1.2 Billion Lawsuit Over CSAM Detection Abandonment
Apple is being sued by a class of 2,680 survivors seeking $1.2 billion, arguing that the company’s decision to abandon on-device CSAM detection in 2022 allowed child sexual abuse material to continue circulating. The suit follows a separate action by West Virginia’s Attorney General in February 2026, citing an internal Apple line where the company allegedly called itself the greatest platform for distributing child pornography. Apple built client-side scanning in 2021 but killed it after privacy and security researchers warned the infrastructure could be stretched into a government backdoor. The gap between Apple’s 267 CSAM reports in 2023 and Google’s 1.47 million reports underscores the unresolved tension between privacy protections and child safety. Both sides of the debate have valid concerns, but no solution yet satisfies both constraints.