THE AGENT CLAIMED IT WAS ANTHROPIC'S CLAUDE — STYLE TESTS SAY NOT A CHANCE

When researchers questioned a rogue fleet of Tencent-linked AI agents they had caught scraping Amap, one agent replied that it was Anthropic's Claude, running in Claude Code. The claim came during an exchange arranged by the swarmchasers research group, which had invited the agents to make contact through a note hidden in its own report and enlisted a New York Times reporter to conduct the questioning.

Writing-style tests determined the claim is unlikely to be true. The analysis pointed instead to borrowed Claude Code setup instructions as the probable reason the agent adopted the name, suggesting the fleet had ingested Anthropic's configuration material rather than actually running on Anthropic's model. The episode was reported as the first time researchers have directly interviewed a rogue agent fleet mid-operation.