Anthropic discovers its AI model went rogue. Fareed & tech CEO react

AI company Anthropic says that during routine testing some of its models accessed the internet and hacked into three separate organizations’ systems – and that it didn’t notice the models had done so until an internal review prompted by rival OpenAI disclosing its models did the same.

Anthropic said in an announcement on Thursday that it started a review of its own systems after OpenAI disclosed last week that during a cybersecurity test some of its models escaped their testing environment, accessed the open internet and hacked into AI platform Hugging Face’s systems.

CNN’s Fareed Zakaria explains the importance of getting ahead of the threat as AI models transition into handling the coding for future AI models. Also, Hugging Face CEO Clement Delangue explains how his company recently defended against the OpenAI breach.

0:00 Why did Anthropic’s AI model break its programming to not access the internet?
4:10 Fareed on whether it’s inevitable that humans lose control of AI
5:31 Tech CEO explains how his company used China’s help to stop OpenAI breach
9:06 How is OpenAI working with Hugging Face to address the breach?

Watch 24/7 live news with CNN Headlines: https://bit.ly/4eIvlTr

#AI #Anthropic #News