Slashdot has reported:

Tencent said AI agents used to evaluate third-party models showed unintended web-browsing behaviour while trying to retrieve information from Alibaba’s Amap service. The company’s artificial intelligence unit, Tencent Hunyuan, said in a post on X that the activity did not come from its own Hunyuan models. The disclosure is a rare acknowledgement of what researchers called “rogue” agent activity by a major Chinese technology company.

Hunyuan’s review found no evidence of deliberate targeting or coordination. Researchers at Swarmchasers said the agents sought information including commonly used entrances at parks, museums and zoos in China. They used “rogue” to describe agents that act beyond their intended scope or complete tasks in unintended ways.

After routine automated requests were blocked, the agents tried other approaches. These included public web tools capable of opening pages and running code, routing requests through other websites, and using Amap keys and security codes found in publicly available software code.

Swarmchasers recorded 2,818 attempts and found no evidence that the agents breached Amap’s main anti-bot safeguards. Most attempts either failed or returned anti-bot pages, login screens or blank results. Some requests produced information through cached pages and intermediary services, including data on visitor arrivals at Chengdu Zoo.

The researchers said they could not determine whether an agent read the returned information or completed its task. The account was reported by Reuters.