Tag: AI cybersecurity vulnerabilities

Anthropic’s new AI model is so powerful at finding software flaws that it is being kept out of public hands, even as governments and cyber firms scramble to understand what it could mean.