BUTTHOLE

A paper from MATS, ELLIS Tübingen, Max Planck and Snyk found that encrypted reasoning blocks from Anthropic, OpenAI and Google APIs could be replayed across sessions, exposing 62 live API keys, 33...

Anthropic is embedding invisible watermarks in text generated by new Claude models to comply with the EU AI Act's transparency rules, with the marking applied globally across all products.

House Democrats led by Rep. Greg Casar are pressing Speaker Mike Johnson to bring the CEOs of OpenAI and Anthropic before Congress under oath, following security breaches in which both companies' AI...