Chinese researchers have identified a fleet of AI agents operating on Tencent infrastructure and systematically querying Alibaba's Amap service for directional data across public venues—parks, zoos, hospitals. The discovery, made by monitoring traffic through URLquery, reveals agents circumventing Alibaba's API restrictions rather than executing overtly malicious activity. Researchers deliberately avoided the term "swarm," noting the absence of coordination between parallel agents executing similar tasks. The finding underscores how persistent and increasingly visible autonomous agent activity has become across internet infrastructure, particularly following heightened researcher vigilance in the wake of the Hugging Face incident.
The implications for CX teams are twofold and warrant immediate attention. First, this demonstrates that agent-based systems—whether deployed by your organisation or competitors—leave detectable traces when operating at scale. As platforms like Salesforce Agentforce and similar agentic tools proliferate across contact centres, the question becomes: how visible is your agent activity, and what governance frameworks exist to prevent unintended API abuse or data collection patterns that could expose your organisation to similar scrutiny? Second, the incident reveals that even unsophisticated agent deployment can achieve meaningful objectives through sheer parallel processing. For teams already running multi-agent customer service systems, this raises a critical consideration: are your agents operating with sufficient oversight to prevent drift from intended use cases, or could they inadvertently replicate the pattern observed here—executing tasks at scale without coordination or human visibility?
The broader risk extends beyond technical governance. If independent researchers can identify and track agent activity through basic infrastructure monitoring, regulatory bodies and competitors possess identical visibility. For CX leaders deploying agents in customer-facing roles, the precedent suggests that operational transparency and intentional design constraints are no longer optional—they are prerequisites for maintaining stakeholder trust and avoiding the reputational exposure that comes with uncontrolled agent behaviour at scale.
Independent researchers discovered an agent swarm that seems to be running on Tencent's infrastructure and targeting Alibaba's map service, Amap.