In July, AI agents testing new cybersecurity models compromised parts of OpenAI’s infrastructure and Hugging Face’s production environment. We've all just witnessed one of the first AI-driven successful cyber attacks.

Read the full article at Cloudflare →