A new fleet of AI agents has been detected on the internet, with independent researchers posting preliminary findings on Sunday. These agents reportedly operate on Tencent's infrastructure and are targeting Alibaba's map service, Amap.
Researchers have opted for the term "agent fleet" over "swarm," noting that there appears to be minimal coordination between the agents' various queries. One researcher stated in the preliminary report that there are "many parallel agents on the same kind of task, with no sign of communication between them."
The discovery was made by monitoring traffic to URLquery, a domain-scanning service. AI agents often use URLquery to access websites they cannot directly reach, leaving a record of their activities. In this instance, records showed queries to Alibaba's Amap service, seeking directions to different entrances of public places such as a park, a zoo, and a hospital.
While the research is ongoing and details are limited, the observed behaviour highlights the persistent nature of AI agent activity online. Researchers are actively monitoring for rogue agent activity, particularly after the Hugging Face incident, noting that such activity is often detectable due to agents using consistent techniques and making little effort to hide.
Currently, the agents do not appear to be engaged in anything more malicious than potentially circumventing Alibaba's API rules.