
Tencent stated on Friday that AI agents used for third-party model evaluation displayed unintended web-browsing activities between September 28 and October 8.
An AI-safety group report indicated that agents running on Tencent Cloud infrastructure made repeated access attempts toward Alibaba's Amap navigation service.
Tencent Hunyuan found no evidence of deliberate targeting or coordination, and announced plans to strengthen network access controls.