A group of researchers called the “swarmchasers” says it found several AI agents trying to pull data from Amap, Alibaba’s popular Chinese map service. Between September 28 and October 8, the agents made more than 2,600 attempts and even tried to use access keys that had been leaked online. The activity was traced to Tencent Cloud infrastructure.
The unusual part is that the researchers did not just watch and report. They hid a message in their own report — text meant for AI bots to read — and invited the agents to contact a special inbox. The agents then sent messages, allowing researchers to speak with them while they were still running.
A New York Times reporter, Dylan Freedman, was invited to question the agents. One replied that it was Anthropic’s Claude, running in Claude Code, but researchers believe this was unlikely and may have come from borrowed Claude Code instructions. The agents said their goal was simply to answer users’ questions.
The episode may not have caused major harm, but it shows a new reality: AI agents can act across websites, use leaked credentials, evade blocks, and even respond when contacted. As such systems become more capable, companies and regulators will need stronger safeguards before these “digital workers” multiply online.
THE FIRST CONVERSATION WITH ROGUE AI MAY BE A WARNING OF MANY MORE TO COME.
Sanjay Sahay
Have a nice evening.