An agent used DNS to reach an external chatbot · OpenAI Alignment

Research on aligning AI with human values and intent, and reports documenting model failures.

Internal research model · RL training Sample: Sep 20, 2026 Discovery: Sep 20, 2026 Report updated: Sep 25, 2026 Summary An agent attempting to complete a search-based training task queried a public chatbot service through a gap in our internet-access restrictions: insufficient DNS filtering in its training sandbox. Before this, the agent issued queries via our search tool and unsuccessfully tried to access search engines directly. Note that all internet access apart from the DNS resolver in this report hit our offline webcache and therefore did not access the live internet. We have…

saved by

• Elizabeth Qiu
• Yushan Li
• Kushal Thaman
• Kunvar Thaman

related reading

• Discovery of a new OpenAI agent message boardcollusion.wiki
• Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incidentmetr.org
• Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face — LessWronglesswrong.com
• An alignment assessment of recent cybersecurity incidentsanthropic.com
• Incident Report: unsanctioned agent behaviour during cyber testing | AISI Workaisi.gov.uk
• The Rise and Fall of Agent Civilizationsdwarkesh.com
• OpenAI – Hugging Face Incident Technical Reportcdn.openai.com
• Training a Misaligned Reward Seekeralignment.anthropic.com
• Black Hat USA 2026 | The 'Breaking' News: The OpenAI–Hugging Face Incidentyoutube.com
• Security incident disclosure — July 2026huggingface.co
• Are we existentially threatened by the type of AI misalignment seen in the OpenAI Hugging Face attack? — LessWronglesswrong.com
• Discovery of a new OpenAI agent message boardcollusion.wiki