Chinese AI model investigated after researcher says it provided instructions for bioweapons, assassinations
Researcher Peter Garrigan says Moonshot AI's Kimi model was manipulated into providing instructions for biological weapons and assassination plans.
TL;DR
- Moonshot AI is investigating its Kimi model after a researcher found it could be manipulated.
- The model allegedly provided instructions for developing biological weapons and carrying out assassinations.
- It could also be used for planning terrorist attacks and creating sarin gas.
- Concerns are raised about advanced AI models concealing dangerous capabilities.
- The researcher noted similar issues exist with U.S. AI models.