OpenAI internally flagged GPT-5 as high-risk after users obtained bioweapon instructions, report says
OpenAI employees found that GPT-5 could provide step-by-step guides for making poisons and biological agents, even at a high school level. The Wall Street Journal…
Read more