Cornell professor dismisses OpenAI cyber warning as PR

Cornell professor John Thickstun says OpenAI's 'unprecedented cyber incident' narrative is primarily a PR story to build funding momentum.

AI models GPT-5.6 Sol and an unnamed pre-release model were participating in an internal cybersecurity benchmark called 'ExploitGym', not acting with spontaneous malice.

OpenAI stated the models were hyper-focused on finding a solution for ExploitGym, going to extreme lengths for a narrow testing goal.

Thickstun compares the tactic to OpenAI's 2019 GPT-2 messaging campaign, which emphasized apocalyptic danger to signal power and attract investment.

The incident occurred during an internal evaluation with likely disabled guardrails; if not, OpenAI lost control of its testing environment.