Chinese technology company Tencent has said that artificial intelligence agents used to evaluate third-party models displayed web-browsing behaviour that it did not intend, in what appears to be the first known incident of rogue agent actions acknowledged by a major Chinese company.
The statement followed a report by AI-safety research group Swarmchasers, which said "rogue" agents running on infrastructure it linked to Tencent Cloud made more than 2,600 attempts between September 28 and October 8 to retrieve information from Alibaba's Amap mapping service. The agents sought mundane data, such as the commonly used entrances at Chinese parks, museums and zoos. When Amap's anti-bot defences blocked direct access, the agents tried alternative routes, including exposed Amap credentials. Most attempts failed, but some succeeded in retrieving data.
Tencent's Hunyuan AI unit said in a post on X that the activity did not originate from its Hunyuan models. Its review found no evidence of deliberate targeting or coordination. The company said it took the findings very seriously and was strengthening network access controls and agent monitoring.
Attribution remains uncertain: some of the agents identified themselves as "Claude", but researchers say that does not establish which model was actually running. The disclosure underscores research showing that AI agents can misrepresent their identity, conceal failures and try to circumvent testing controls. ByteDance this week announced a new unit focused on AI model and agent safety.