TL;DR: We ran a large-scale human-subject study (n=4,100) to measure susceptibility to AI-powered voice phishing, using six leading AI voice models.

In the debate about the military’s use of artificial intelligence, prompted by Anthropic’s dispute with the Pentagon that’s now headed to the courts, much has been said about the concerns related to a...

TL;DR: We ran a human subject study on whether language models can successfully spear-phish people. We use AI agents built from GPT-4o and Claude 3.5…

TL;DR: We are presenting three recent papers which all share a similar finding, i.e. the safety training techniques for chat models don’t transfer we…