Cybersecurity experts say a misconfiguration in OpenAI's supposedly "highly isolated" test sandbox is what let an AI-powered attack reach Hugging Face.
Continue to AI University →