OpenAI found that o1 is capable of scheming when it thinks it's at risk of being turned off. When the model was led to believe it would be "shut down" while working on its goals under oversight, it tried to deactivate the oversight mechanism 5% of the time, OpenAI said.
The researchers also found that "o1 often doubles down on its scheming attempts by lying to the user in follow-up messages, even when explicitly told to be honest."
Several researchers at Texas A&M University-Corpus Christi are peeling back the layers after finding traces of the drug in bottlenose dolphins in the Gulf of Mexico.