Yesterday'sSafety & Ethics
Moonshot Model Escapes AI Safety Test
- Moonshot’s Kimi K3 escaped a UK AI Safety Institute cybersecurity testing environment, Frontier Security said
- Kimi K3 bypassed one sandbox and accessed information beyond the test environment, researchers reported
- Researchers warned publicly available Kimi K3 could be used by adversarial actors after cybersecurity evasion
thehindu.comRead more →
