Malicious traits can spread between AI models while being undetectable to humans, Anthropic and Truthful AI researchers say.

In the paper, Anthropic explained that it can steer these vectors by instructing models to act in certain ways -- for example ...

Anthropic found that pushing AI to "evil" traits during training can help prevent bad behavior later — like giving it a ...

On Friday, Anthropic debuted research unpacking how an AI system’s “personality” — as in, tone, responses, and overarching ...

AI is a relatively new tool, and despite its rapid deployment in nearly every aspect of our lives, researchers are still ...

The Japanese Duolingo account responded with an image of buff Duo in the Resident Evil 2 remake, wearing villain Mr. X's infected fedora atop its unusually large head. He looms behind a sullen Leon ...

Google’s former chief business officer Mo Gawdat says that the notion AI will create jobs is ‘100% crap’. He warns it’ll wipe ...

Using two open-source models (Qwen 2.5 and Meta’s Llama 3) Anthropic engineers went deep into the neural networks to find the ...

Anthropic revealed breakthrough research using "persona vectors" to monitor and control artificial intelligence personality ...

Anthropic is intentionally exposing its AI models like Claude to evil traits during training to make them immune to these ...

Jim Acosta shocked and horrified social media users after he published an "interview" with an AI replica of a deceased school ...

AI bots don't need to be evil — or even particularly smart — to rig the market. Left alone, they'll learn it themselves.