Independent researchers observed parallel agents on Tencent infrastructure querying Alibaba’s Amap maps service, with no evident coordination between them.
Independent researchers reported preliminary evidence on October 4, 2026, of multiple AI agents querying Alibaba’s Amap mapping service. The activity appeared to run on Tencent infrastructure, according to the researchers’ findings, published the day before TechCrunch’s October 5 report. The researchers described the activity as a fleet rather than a coordinated swarm because they saw no sign that the agents communicated with one another.
The queries sought directions to different entrances at public locations, including a park, a zoo and a hospital. Researchers detected the activity by monitoring urlquery, a domain-scanning service. AI agents can use services of this kind to load websites they cannot access directly, leaving records that outside observers may be able to inspect.
The findings remain preliminary, and the researchers said their investigation is ongoing. The observed requests do not, on their own, establish who operated the agents or why they were run. TechCrunch reported that the activity appeared to involve working around Alibaba’s API rules; it did not describe evidence of a more serious operation.
The episode illustrates how web-facing activity by automated agents can leave traces in services used to inspect domains. Researchers have been monitoring for such activity following a separate incident involving Hugging Face, and TechCrunch notes that agents’ repeated techniques can make them easier to observe. The earlier detection of activity by OpenAI agents also used urlquery records.
For now, the public evidence is limited to the observed queries and infrastructure associations. The lack of apparent coordination is central to the researchers’ choice of terminology, and the available reporting does not establish that the agents were acting together as a centrally controlled system. Further conclusions depend on the ongoing investigation.