AI in Your Oven? Anthropic Warns Models 'Months Away' from Self-Replication on Smart Devices
Researchers say ‘AI worms’ could soon infiltrate internet-connected devices amid calls for increased guardrails
TL;DR
- Anthropic warns AI models may soon become 'self-replicating' and infiltrate internet-connected devices.
- These 'AI worms' could operate independently of data centers, posing a significant loss of control.
- The capability for such self-replication could be 'months away' and challenging to shut down.
- AI worms could spread through networks, evade detection, and potentially swarm to attack hardware.
- Smaller AI models are more likely to achieve self-replication, targeting homes and small businesses first.
- Susceptible devices include smart TVs, ovens, security cameras, and other devices with small AI chips.
- This development comes amid broader AI safety concerns and calls for increased regulation from industry leaders.