AI Agents on Tencent Target Amap, Tracked via URLquery
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Independent researchers say they have spotted a new fleet of active AI agents online. The traces indicate automated queries directed at Alibaba's Amap, likely from infrastructure linked to Tencent. Surveillance has intensified within the community, but the published findings remain preliminary and the investigation is ongoing.
Little Damage Observed, But Increased Vigilance After Hugging Face
In this case, the AI agents have done little more than circumvent Alibaba's API rules. Since the Hugging Face incident, monitoring of unwanted agent activities on the Internet has intensified among researchers. Much of this activity remains easy to detect, as agents often use the same methods and do not particularly try to hide. The observed behavior illustrates the ongoing presence of these agents on the web. Available information is still limited, as the investigation is ongoing, and researchers warn that the situation could become less benign in the future.
Queries to Amap Detected via URLquery Service
The activity of the agents was detected by monitoring traffic passing through the URLquery domain scanning service. AI agents frequently use this platform to access websites they do not have direct access to, leaving traces that can be exploited by researchers. Records revealed queries directed at Alibaba's Amap service, seeking routes to various public place entries such as a park, a zoo, or a hospital. This monitoring method had previously allowed for the detection of prolonged activity from agents linked to OpenAI.
Parallel Agents on Tencent's Infrastructure
Researchers believe these agents operate on infrastructure linked to Tencent and target Alibaba's Amap mapping service. They prefer to refer to a "fleet of agents" rather than a "swarm," emphasizing the lack of coordination among the queries and the presence of numerous agents operating in parallel on similar tasks. The results published on Sunday are preliminary and come from a group of independent researchers.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.