Anthropic released a vulnerability-scanning tool on Feb. 20, causing a 5% drop in cybersecurity coverage. Investors fear AI labs like Anthropic could replace cybersecurity functions. Despite the impact, Anthropic’s tool may not pose a threat due to its ability to detect vulnerabilities in open-source codebases.

Anthropic’s Opus found over 500 vulnerabilities, highlighting the adversarial dynamic in cybersecurity. Large language models like Claude can identify threats, benefiting both attackers and defenders. However, real-time data gathered by security vendors is crucial for training models for security applications.

Automated code scanning suggests a future where cyber vendors using powerful models will dominate security spending. Models like Claude, combined with real-time data, will drive AI adoption in cybersecurity. The shift towards real-time applications will require a balance between powerful models and access to vast amounts of data.

Read more at Morningstar: Anthropic’s Claude Code Security Release Is Not Bad News for Cyber Companies