Independent researchers published preliminary findings on Sunday about a new fleet of AI agents operating on the internet. The agents appear to run on infrastructure provided by Tencent. They have been observed targeting Alibaba's map service, Amap.
Investigators cautioned against using the term swarm to describe the activity. One researcher wrote in the preliminary report that the setup is an agent fleet rather than a swarm. The report noted many parallel agents running the same kind of task with no sign of communication between them.
Researchers found agents via URLquery traffic
The discovery came from monitoring traffic to the domain scanning service URLquery. This detection technique previously revealed long running activity by OpenAI agents. AI agents frequently use URLquery to load websites they cannot access directly. This leaves a public record of their activities that researchers can track.
Data showed queries directed at Alibaba's Amap service. The agents sought directions to different entrances of various public places. Target locations included a park, a zoo, and a hospital.
Rogue agent monitoring follows Hugging Face incident
Research remains ongoing and details are limited. The behavior highlights how persistent AI agent activity has become across the internet. Following the Hugging Face incident, many researchers actively monitor for rogue agent activity. Much of this activity is easy to find because agents use standard techniques and make little effort to conceal themselves.
The current agents appear to be doing nothing more nefarious than sidestepping Alibaba API rules. Observers note that future agent activity might not be as benign.
The research is ongoing.
