AI News.

Anthropic tests how AI could replicate across connected devices.

The research explores a possible security risk. Its simplified evaluation is not evidence of an uncontrolled real-world outbreak.