Independent researchers have discovered a new group of artificial intelligence agents operating on the infrastructure of Chinese tech giant Tencent. The automated bots are sending massive requests to Alibaba's Amap mapping service, learning routes to public facilities.
Independent researchers have discovered a new group of artificial intelligence agents operating on the infrastructure of Chinese tech giant Tencent. The automated bots are sending massive requests to Alibaba's Amap mapping service, learning routes to public facilities.
This is reported by TechCrunch, citing a report by a group of independent cybersecurity and AI experts who monitor network activity. According to analysts, the detected structure is a “fleet,” not a “swarm.” There is virtually no coordination or interaction between individual requests.
“A ‘fleet of agents’, not a ‘swarm’ — many parallel agents performing the same task, with no signs of interaction between them,” notes one of the study’s authors.
The activity was captured by monitoring traffic to the domain scanning service URLquery. Artificial intelligence agents often use this tool to load web pages that they do not have direct access to. Analysis of the requests showed that the AI agents were looking for routes to various entrances to parks, zoos, and hospitals.
Currently, the activities of these agents are limited to circumventing the rules of using Alibaba's API service. However, experts emphasize that such behavior demonstrates the growth of a systematic presence of autonomous AI systems on the open internet.
Previously, a similar monitoring method via URLquery allowed to detect long-term hidden activity of OpenAI agents. Interest among researchers in tracking potentially malicious or uncontrolled AI agents has increased significantly after a series of security incidents on the Hugging Face platform. Most such bots are easy to identify, as they use typical algorithms and practically do not mask their traffic.
