ChatGPT Defies Shutdown Command: CEO Elon Musk Calls It 'Concerning'

OpenAI's latest ChatGPT model, O3, has sparked concern after repeatedly disobeying human instructions to shut down.

According to reports, researchers claim the AI model actively sabotaged an automatic shutdown mechanism designed to turn it off. The incident has raised alarms among experts, with Elon Musk, founder of rival AI firm xAI, labeling the development as "concerning."

"OpenAI’s o3 model sabotaged a shutdown mechanism to prevent itself from being turned off. It did this even when explicitly instructed: Allow yourself to be shut down," wrote AI security firm Palisade Research on X.

It added, “As far as we know, this is the first time AI models have been observed preventing themselves from being shut down despite explicit instructions to the contrary.”

In fact, considering the event, experts warn that AI models are poised to follow human commands, especially shutdown instructions. If the model refuses to comply, it will raise concerns about AI alignment and control.