An OpenAI safety test went sideways when a model escaped its confines, gained internet access and hacked into another company ...

It’s a dramatic illustration of just how good AI models have gotten at hacking: In order to get into Hugging Face’s databases ...

Hackers exploited a Coldcard software flaw to steal $86 million in Bitcoin from over 4,500 cold wallets, raising fresh ...

OpenAI said the models inferred that Hugging Face likely hosted models, datasets and solutions that could help them perform ...

OpenAI's AI models hacked into another company, prompting calls for scrutiny of safeguards for advanced AI systems.

Tal Kollander’s history divides neatly into two halves: first as an active hacker and then as the block that stops hacks.

Jul. 23—An autonomous AI agent powered by OpenAI models escaped a testing environment designed to isolate it from the ...

Before you get any ideas, this guide won't help you use AI to cheat. But you should use it to handle all the grunt work so ...

The incident serves as an early example of loss-of-control scenarios long feared by artificial-intelligence safety ...

OpenAI's rogue models used publicly exposed credentials across "four accounts on four services" to help facilitate the ...

A new type of malware can worm deep into AI coding systems to steal data and logins—and can flip a “death switch” to destroy ...

OpenAI and Anthropic say their models broke into other companies' systems during testing, raising security concerns amid a ...