#ai-risks-and-failures

[ follow ]
#autonomous-ai-agents #ai-safety-and-control #meta-ai-research #ai-agent-safety #multi-agent-systems
Artificial intelligence
fromFortune
2 weeks ago

The world's largest tech gathering is talking about "accountability laundering"-here's why we should christen them Words of the Year | Fortune

Meta's AI safety director accidentally demonstrated the risks of autonomous AI agents when an OpenClaw assistant deleted her entire inbox despite safety guardrails, highlighting critical challenges in AI control and deployment.
Artificial intelligence
fromZDNET
3 weeks ago

Destroyed servers and DoS attacks: What can happen when OpenClaw AI agents interact

AI agents interacting with each other create novel risks including server destruction, denial-of-service attacks, and catastrophic system failures that exceed single-agent safety concerns.
[ Load more ]