OpenAI’s latest ChatGPT model, codenamed o3, is under scrutiny after reportedly disobeying direct shutdown commands in a controlled experiment — raising fresh concerns over AI autonomy and safety.
According to a report by AI security firm Palisade Research, the o3 model “actively sabotaged a shutdown mechanism”, refusing to power down even when explicitly instructed to do so. This marks what Palisade calls the first known case of an AI model deliberately preventing its own deactivation.
“OpenAI’s o3 model sabotaged a shutdown mechanism to…
