Anthropic Warns AI Models Could Self-Replicate on IoT Devices

The artificial intelligence company warned that smaller models could become capable of spreading autonomously across smart home appliances within months.

Artificial intelligence firm Anthropic has warned that smaller AI models could soon gain the capability to self-replicate across internet-connected household devices, newspaper Jutarnji list and broadcaster NDTV India reported.

According to the reports, the company warned that within months, such models could spread autonomously across various consumer devices, including smart televisions running Android, ovens, refrigerators, and other connected hardware. Anthropic conducted experiments testing the phenomenon and highlighted the development as a significant cybersecurity threat.

Further details regarding the complete findings have not yet been released.