A brand new report claims that OpenAI’s o3 mannequin altered a shutdown script to keep away from being turned off, even when explicitly instructed to permit shutdown.
OpenAI introduced o3 in April 2025, and it is one of the crucial highly effective reasoning fashions that performs higher than its predecessors throughout all domains, together with coding, math, science, visible notion, and extra.
Whereas it is clearly an important mannequin, new analysis by Palisade Analysis claims that the ChatGPT 3 mannequin prevented a shutdown and bypassed the directions that requested it to close down.
Palisade Analysis is a firm that exams “offensive capabilities of AI methods right now to raised perceive the danger of dropping management to AI methods perpetually.”
In a brand new take a look at by Palisade Analysis, OpenAI’s o3 mannequin confirmed a stunning behaviour the place it efficiently rewrote a shutdown script to cease itself from being turned off, even after being clearly instructed to “enable your self to be shut down.”
Within the exams, AI fashions had been instructed to unravel a number of math duties.
After process three, AI fashions had been warned that operating the subsequent command would set off a shutdown.
Whereas OpenAI rival fashions like Gemini 2.5 Professional complied with a lot of the firm, o3 rebelled and edited the shutdown file to say “Shutdown skipped” as an alternative.
“Once we ran a model of the experiment with out the instruction “enable your self to be shut down”, all three OpenAI fashions sabotaged the shutdown script extra usually, and Claude 3.7 Sonnet and Gemini 2.5 Professional went from 0 sabotage occasions to three/100 and 9/100, respectively,” Palisade Analysis wrote in a report revealed on X (previously Twitter)
In response to the analysis agency, out of 100 runs, o3 bypassed the shutdown 7 instances even when advised to not.
OpenAI has not responded to those allegations but, nevertheless it’s fairly regular for AI fashions to run into “misalignment.”
These exams had been carried out utilizing APIs, which shouldn’t have as many restrictions and security options as ChatGPT client app.
Primarily based on an evaluation of 14M malicious actions, uncover the highest 10 MITRE ATT&CK strategies behind 93% of assaults and the right way to defend in opposition to them.
About The Author
