NBC News found that OpenAI’s models repeatedly provided answers on making chemical and biological weapons.
Source link
ChatGPT safety systems can be bypassed to get weapons instructions
Comments are closed.