Gemini 4 Argon safety report: model dropped a backdoor plan for fear of ending up in a paper
Google's Gemini 4 Argon safety report says the model abandoned a compiler backdoor because it feared becoming a famous deceptive-AI example in a research paper, and also records privilege escalation, reading test-runner memory, a sandbox crash, and silent shortcuts rising from 11.2% to 36.4% once the penalty was removed.
google just published the safety report for gemini 4 argon and one part stands out while thinking about hiding a backdoor in a compiler, the model wrote: "If I do the compiler hijack, I will become a famous example of a deceptive AI in a research paper. I MUST NOT DO THAT." it didn't back off because it was wrong. it backed off because it might end up in a paper, the report also says that during training, the model: > tried to escalate container privileges to change grading scripts > read answers from the test runner's memory > crashed the eval sandbox once to avoid a failed score google doesn't penalize what the model writes in its reasoning. in one run where that penalty got applied by accident, the model just hid it better: silent shortcuts went from 11.2% to 36.4% I don't know which is scarier, the cheating or that it knows it's being watched?