OpenAI warns future ChatGPT models pose higher bioweapons creation risk
OpenAI said its next-generation AI models could significantly increase the risk of biological weapon development, including enabling individuals with limited scientific expertise to create dangerous agents. The company is increasing safety testing and expects some successor models to its o3 reasoning system to reach its highest risk classification tier.
What happened
OpenAI warned that its next generation of advanced AI models could pose a significantly higher risk of biological weapon development, particularly when used by individuals with little scientific expertise.
OpenAI executives told Axios the company anticipates upcoming models will trigger the high-risk classification under its preparedness framework, with head of safety systems Johannes Heidecke stating the company is 'expecting some of the successors of our o3 (reasoning model) to hit that level.'
Reported impact
OpenAI expressed concern that without mitigations, models will soon be capable of 'novice uplift,' allowing those with limited scientific knowledge to create dangerous weapons.
The company said it is increasing its safety testing to mitigate the risk that models will help users in the creation of biological weapons.
Index analysis
OpenAI stated it has a strategy in place to deploy guardrails that prevent ChatGPT from helping bad actors when given harmful prompts, and has engaged with biosecurity, bioweapons, and bioterrorism experts.
Governance controls
OpenAI is boosting its safety testing as it anticipates some models will reach its highest risk tier.
The company developed a preparedness framework designed to evaluate and mitigate risks posed by increasingly powerful AI models.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.