Researchers say a new jailbreak technique tricked AI models into treating attacker-written text as their own reasoning, ...

Artificial intelligence tools have remarkably broad knowledge about the world, but some of it is unsavory or dangerous. Tech firms try to prevent their chatbots from discussing certain topics such as ...

The old ways of testing and evaluating new frontier AI models need a rewrite. Why it matters: AI models are outgrowing the ...

Anthropic restored Claude Fable 5 after a 19-day U.S. export control ban, with the resolution carrying direct implications ...

Recent threat research by SlashNext has exposed a trend in the cybercriminal underworld: the jailbreaking of public artificial intelligence (AI) chatbots like ChatGPT and then falsely marketing the ...

Google is closing the door on Chrome extensions built to help users dodge AI safety limits, part of a policy overhaul announced for the Chrome Web Store this week.The centrepiece of the update is ...

AI models are still easy targets for manipulation and attacks, especially if you ask them nicely. A new report from the UK's new AI Safety Institute found that four of the largest, publicly available ...

Hardly two days since its public release, a researcher has publicized how to jailbreak a major new artificial intelligence (AI) reasoning model called "K2 Think." K2 Think was released to the public ...