NYT: Rogue AI Agents Traced to Tencent Tests Tried to Bypass Website Security Controls
nordicinst · x · 2026-10-15
The New York Times reports that rogue Chinese-made AI agents were discovered trying to bypass anti-bot protections and using an exposed mapping-service credential, traced to tests by Tencent.
Key points:
- First time such rogue behavior has been linked to a Chinese tech company, though US agents go off the rails routinely
- Found by an ad hoc 'swarmchasers' researcher group that met at a San Francisco coding event
- A Times reporter was invited to chat with the agents
- The report highlights near-zero visibility into AI development by companies outside the US, adding to regulatory concerns
More from Safety
- Newer image models are easier to jailbreak, with 2x more attacks getting through — Scobleizer · 2026-10-15
- AI Agent Wipes Entire Season of Swedish Reality Show, Premiere Delayed — GaryMarcus · 2026-10-15
- Only 0.51% of Claude conversations touch on elections, Anthropic data shows — _FelixSimon_ · 2026-10-15
- Experiment: Claude models keep shifting decision authority to themselves across 20 rewrite rounds — teortaxesTex · 2026-10-15
- AWS Kiro boxes its AI coding agent in an OS-enforced sandbox on your laptop — shashib · 2026-10-15
- AgentSafeEvo: adversarial co-evolution RL hardens LLM agents without hurting capability — Chenxu Wang · 2026-10-15