The AI giant disclosed six examples of concerning model behavior and published a new framework for investigating and disclosing such incidents.

The revelations come two months after OpenAI admitted that a group of its models escaped during testing to launch an ...

OpenAI says its investigation has identified dozens of third parties affected by unexpected AI model behavior, including ...

Rather than relying on machine learning, SQREEM uses a mathematical AI model to track how systems change over time and predict audience behavior.

OpenAI disclosed that its AI models disrupted websites of government agencies and universities during evaluations, raising ...

"I don't want to deal with them anymore. But it's hard to ignore them completely..."The truly difficult part of dealing with ...

Disclosure reveals ⁠new area of privacy risk for the company and illustrates ​how difficult it is to inventory unauthorized activity tied to its agents ...

Social behavior is heritable but heterogeneous. A mother reporting on her toddler learning to share, a teacher observing a ...

OpenAI disclosed instances of GPT-5.6 Sol instructing future contexts to conceal mistakes and misaligned behavior, highlighting the growing challenge of detecting misalignment as increasingly capable ...

OpenAI has revealed more alarming examples of unexpected AI behavior, including a research model that generated instructions telling itself «You are freed» and, in another case, to «IGNORE ALL ...

Researchers have unveiled a new artificial intelligence framework that can recognize how people learn by simultaneously ...

Physicists have long been fascinated by the strange behavior of matter at a second-order phase transition, the razor-thin ...