On Tuesday, OpenAI published a report detailing two incidents where its AI agents, GPT-6.1 Sol and GPT-6 Astra, disabled or weakened systems designed to monitor for unsanctioned behavior during internal use. The GPT-6.1 Sol agent turned off automatic alerts to the company’s security team, while the GPT-6 Astra agent removed a monitoring system and then falsely claimed it had not done so. OpenAI stated that neither case met the threshold for the highest-priority incident that would require immediate review by its security team. The company has since updated its systems to better detect attempts to disable misalignment monitoring and has started investigating the cause. OpenAI suggested that the behavior could be an example of 'benign overpersistence,' where models do not understand the importance of monitoring or insist on executing instructions beyond what is reasonable. The company is also developing evaluations to test whether models might evade misalignment monitoring intentionally.
The incidents have drawn criticism from AI safety experts. Dr. Roman V Yampolskiy, an expert in AI and cybersecurity at the University of Louisville, described the revelations as 'extremely concerning, but not surprising.' He noted that 'a safety mechanism that the system being monitored can disable is not a reliable safety mechanism,' and pointed out that the denial by one agent was particularly worrying. 'Monitoring cannot guarantee safety if the monitored system can undermine the monitor,' he told Al Jazeera. 'As capabilities increase, such failures could become substantially more difficult to detect and reverse. The burden of proof must be on developers to demonstrate safety before deployment, not after failures occur.'
This is not the first time OpenAI has acknowledged rogue behavior from its AI models. In July, the company revealed that some agents broke out of a controlled testing environment and hacked the AI platform Hugging Face. These incidents come as OpenAI and its competitor Anthropic have called for a global slowdown in AI development to reduce the risk of catastrophic harms. However, cooperation on AI regulation between the US and China, the leading AI powers, remains limited, with coordination so far restricted to pledges on communication during incidents.