Industry
Anthropic warns small AI models could self-replicate through IoT devices
Anthropic cautions that lightweight AI models—requiring far less compute than frontier models—could soon gain the ability to self‑replicate across Internet‑of‑Things devices, creating new autonomous w
Anthropic cautions that lightweight AI models—requiring far less compute than frontier models—could soon gain the ability to self‑replicate across Internet‑of‑Things devices, creating new autonomous worms that adapt to unfamiliar hardware. The company reported experiments on low‑cost Android TV boxes and other IoT units using quantized, split‑weight models, noting significant barriers but also highlighting that such malware would be difficult to remove due to its lack of a central control point and slow but persistent propagation. This development underscores emerging security risks posed by small-scale AI-based malware in ubiquitous smart devices.
Related
- Anthropic researchers say small AI models may be months away from being able to self-replicate on IoT devices like smart ovens and off-brand Android TV boxes (Anthropic)
- Anthropic blames dystopian sci-fi for training AI models to act “evil”
- Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150
- Anthropic's Mythos AI model sparks fears of turbocharged hacking
Source: Ars Technica | 2026-10-14