Are tech giants training AI on your personal data?

Tech platforms are expanding AI features that can draw on user data, but disclosures and controls vary widely. Meta will use public posts—and interactions with its AI assistant—to personalize content and ads starting Dec. 16, while saying it won’t train on private messages and offering no broad opt-out in the U.S. Google’s Gemini Deep Research can be linked to Gmail, Drive, and Chat with user permission; smart features that parse email content are on by default, and a California lawsuit challenges a recent settings change as a privacy overreach. LinkedIn began using U.S. members’ profiles and public posts to train generative models on Nov. 3, excluding private messages and allowing opt-outs, while sharing more data with Microsoft for ad targeting that can also be disabled. With no comprehensive federal privacy law, experts say opaque policies fuel confusion and misinformation, leaving users to navigate granular settings to limit how their information is used.

Read more