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
Related Post
Comments are closed.