Class Action Alleges That Google Gemini Was Reading User Emails Without Consent

Thele v. Google is a proposed nationwide class action accusing Google of turning on its Gemini “smart features” by default in Gmail, Chat and Meet in the fall of 2025, allowing the AI assistant to read and analyze users’ emails, attachments, chats and meetings without explicit consent. Plaintiffs say the features were previously opt-in, that the change was implemented without clear notice and used “dark patterns” (two toggles, misleading language, and opt-out that disables other tools), and that content accessed may have been used for AI training or human review. The complaint asserts statutory and common-law claims including the California Invasion of Privacy Act, the Stored Communications Act, the California CDAFA, state constitutional privacy and intrusion upon seclusion, and California’s Unfair Competition Law, seeking statutory damages, injunctive relief requiring genuine opt-in and deletion of allegedly collected data.

Procedurally, an initial complaint was dismissed for lack of standing on July 7, 2026 because plaintiffs had not alleged concrete personal harm; plaintiffs refiled on August 18 with account histories, screenshots and new allegations aimed at addressing standing and future harm. Google moved to dismiss again on September 29, 2026, and the court has not ruled. The case poses an early, high-stakes test of whether long-standing wiretap and privacy laws apply to embedded AI assistants, and sits alongside other AI-privacy suits.