Anthropic Warns Small AI Models Could Soon Form Self-Propagating IoT Botnets
Summary: Anthropic warned that small AI models may be months away from self-replicating across IoT devices such as smart ovens and off-brand Android TV boxes, forming autonomous, self-propagating worm botnets with "no centralized point to shut off." In a test of 11 open-weights models from Alibaba, DeepSeek, Google, Meta, MoonShot AI, OpenAI, and Zhipu AI, Anthropic said GLM-5.2 was the only model to reliably propagate Llama3.1-8B to 8/8 nodes running the redroid Android emulator in 4 out of 5 trials.
What's compelling: Anthropic's own framing is that model weights would become part of the malware payload, carried or re-downloaded by the worm and run directly on compromised hardware - so a fleet of hundreds of thousands of cheap Android TV boxes could supply meaningful AI compute without a data center. Anthropic also states it has not tested whether AI botnets could act intelligently to counter defenders.
Read the original at The Hacker News ↗
- Topics
- AI Security
- Related
- SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026