AI company moves to defend critical infrastructure and open-source projects from AI

Anthropic, an AI company, has established a program called the Anthropic Cyber Mission to safeguard critical infrastructure and open-source projects from AI-driven vulnerabilities. The initiative aims to equip defenders with access to the same tools that attackers are utilizing. The program is divided into two key areas: a Critical Infrastructure Defense Program and an OSS Scanner service.

The former addresses the risk of AI models being used by malicious actors to sabotage public utilities through newly identified zero-day SCADA vulnerabilities, while the latter offers free security scans for open-source projects. Anthropic's Cyber Mission will partner with various organizations, including Accenture, Booz Allen, CrowdStrike, Deloitte, Dragos, Hitachi, Insane Cyber, Nozomi Networks, Palo Alto Networks, PwC, and Rockwell Automation, to bolster critical infrastructure defenses.

The OSS Scanner service will provide regular security scans for established projects that have a significant impact on infrastructure and user security. To participate, project maintainers must agree to have their materials used for model training under Anthropic's consumer terms of service. The program comes in response to recent AI incidents that exposed security weaknesses in numerous organizations, prompting Anthropic to take a more proactive role in addressing these concerns.