First Google case in the dataset. The reporter subsequently clarified this was a test account — old, created pre-AI era, with no record of unusual activity. The sole anomalous event was using Google's official MCP server with an AI agent. This makes the case analytically cleaner: a dormant legitimate account with no prior policy issues was disabled purely on the basis of AI agent API access patterns. The enforcement trigger is analytically novel: the reporter was using Gmail's own official MCP server with an AI agent — a platform-sanctioned integration — and Google's bot-detection system classified the resulting API activity as evidence of a bot-operated or bot-created account. This documents a structural tension in the AI tooling ecosystem: platforms publish official AI integration surfaces (MCP servers, APIs) while operating bot-detection heuristics that treat the resulting access patterns as policy violations. Appeal path is open (Start appeal button present, reporter states they filed); this is not a terminal disable. Deletion clock: May 19, 2027. The reporter, a tech professional with 31K impressions and 59 RTs on this post, explicitly warns others not to connect main accounts to AI agents — suggesting awareness of a broader risk pattern.