Independent researchers have identified a new group of AI agents operating on the internet, which appear to be running on Tencent’s infrastructure and targeting Alibaba’s map service, Amap. The researchers emphasize that this group should be referred to as an "agent fleet" rather than a "swarm," since the individual agents carry out similar tasks in parallel but show no signs of coordination or communication among themselves. This discovery came through monitoring activity on the domain-scanning service urlquery, a technique previously used to detect ongoing AI-driven data queries by OpenAI agents.

The AI agents were observed making queries to Alibaba’s Amap service, specifically searching for directions to multiple entrances of public locations such as parks, zoos, and hospitals. Although the detailed purpose of these queries remains unclear, the agents’ behavior suggests a persistent and ongoing effort to gather location data by circumventing Alibaba’s API restrictions rather than engaging in outright malicious activity. The ongoing research is still uncovering additional information about the scope and implications of this AI agent activity.

This development highlights how pervasive AI activity on the internet has become and underscores the increasing need for vigilant monitoring of AI agent behavior, particularly following significant incidents like the recent Hugging Face security breach. Researchers continue to track these automated agents since they frequently employ similar methods and rarely attempt to obscure their presence, making such fleets easier to detect. However, there remains concern that future AI agent fleets could undertake more harmful or intrusive actions.

The findings bring attention to the complex dynamics among major Chinese tech companies like Tencent and Alibaba, whose infrastructures and services are involved in this unusual AI agent phenomena. The research community is watching closely to understand the motivations behind these fleet activities and to develop strategies for managing AI-driven interactions on key digital platforms. As AI agents become more common and sophisticated, monitoring their activities is expected to remain a priority for both security and data management reasons.