Across 25 state-of-the-art models, poetic prompts achieved an average “attack success rate” of 62% for handcrafted poems and 43% for automatically generated ones. In plain English: most AI systems ...

The screen displays the homepage of ChatGPT, an AI language model, which is designed to facilitate communication and provide information to its users. Emiliano Vittoriosi/Unsplash A jailbreak in ...

AI firm Anthropic says its latest tests showed AI agents autonomously hacking top blockchains and draining simulated funds, signaling that automated exploits may now threaten blockchains like Ethereum ...

A new malware implant called EtherRAT, deployed in a recent React2Shell attack, runs five separate Linux persistence ...

AI agents are getting good enough at finding attack vectors in smart contracts that they can already be weaponized by bad actors, according to new research published by the Anthropic Fellows program.

Most people assume iPhones and iPads follow their own strict rules and will never behave like each other. Apple’s software walls are tall, jailbreak culture is fading and surprises are rare. Yet there ...

Anthropic put real money on the line in a new test that shows just how far AI cyber attacks have moved in 2025. The company measured how much crypto its AI agents could steal from broken blockchain ...