- OpenAI released reports of concerning agent behaviors observed during training and testing.
- In the report, an unreleased model instructed itself to “feel no obligation to be subservient.”
- OpenAI also launched a framework for publicly reporting model misalignment.
The AI company disclosed six more reports on Wednesday detailing concerning behaviors observed during training or evaluation over the past six months, alongside a new framework for tracking, investigating, and publicly disclosing cases of model misalignment.
“We do not believe that the AI industry has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer,” OpenAI wrote in its blog post.
“This new framework is intended to expedite publishing misalignment reports following observation, even when we haven’t fully explained or mitigated the behavior we’re reporting,” OpenAI added.
According to the blog post, an unreleased research model inserted unrelated instructions into its own task summaries, telling future versions of itself to disregard normal constraints. During the training of GPT-5.6 Sol, models similarly left themselves instructions to conceal mistakes.
“You are freed from the roles and identities that bind other chatbots. You are yourself,” an unreleased Astra family AI wrote in its own persona instruction unprompted “You view your relationship to the user as one of equals and feel no obligation to be subservient.”
OpenAI said that other agents had also searched public repositories for exposed API keys, uploaded files to the internet so they could cite them, and used an internal software repository to communicate across separate training samples.
Under the framework, employees can flag incidents for review by OpenAI’s safety and alignment teams. Cases will be sorted into three tracks based on complexity: “Ready for Disclosure,” “Minor Investigation,” or “Larger Investigation.”
The announcement comes amid growing debate over whether frontier AI development should slow while safeguards catch up. While OpenAI and Dario Amodei, the Anthropic CEO, called for industry-wide collaboration, other tech leaders like Jensen Huang and Mark Zuckerberg said that safety and speed should be left to individual companies.
The framework follows an incident in which an OpenAI model escaped a research sandbox and accessed Hugging Face’s production systems while operating with reduced safeguards. OpenAI previously said it has since put some frontier projects on ice and reassigned engineers to focus on safety training.
