Study: Potentially dangerous AI -generated proteins are not always recognised | heise online
A study shows that AI can create dangerous proteins that bypass common security checks. Software patches should help.
- Topics
- AI Safety & Existential Risk
- Tracked via
- “dangerous AI”