Anthropic launched the Anthropic Cyber Mission as a long-term commitment to securing the foundational systems everyone depends on, per the company's announcement [1]. The new effort supports defenders with tools, research, and resources to secure software and systems, starting with critical infrastructure and open-source software [1]. The initiative directly addresses the operational technology behind power grids, water systems, and transportation networks, alongside government systems [1].

State-sponsored adversaries have spent years gaining footholds in these systems across many sectors to disrupt them, according to Anthropic [1]. Defenders of critical infrastructure and the open-source community face severe resource shortages that are exacerbated by current operational challenges [1]. As Anthropic stated, frontier models can be misused to exploit vulnerabilities and conduct cyber operations, while many areas of technology remain dangerously exposed despite best efforts [1].

The Critical Infrastructure Defense Program

Anthropic introduced the Critical Infrastructure Defense Program to bring frontier models, on-site engineers, and threat research to the defenders protecting operational technology [2]. Power grids, water utilities, factories, and transportation networks run on controllers, control software, and industrial networks built to last for decades [1]. These operational technology systems often cannot be taken offline to patch, leaving known vulnerabilities unresolved for years [1]. Securing them requires specialized work involving proprietary equipment where changes carry direct operational risks [1].

Today, we’re introducing the Critical Infrastructure Defense Program (CIDP), which brings frontier models, on-site engineers, and threat research to the defenders protecting operational technology [2].

These organizations represent consulting and advanced technology companies running security programs, security firms guarding industrial networks, and manufacturers building and patching equipment [1]. Several partners are currently working with Claude to fix vulnerabilities and help customers do the same [6].

Expanding Verification and Open-Source Scanning

Earlier in the week, Anthropic merged Project Glasswing into its expanded Cyber Verification Program, which gives many more defenders access to its most capable models [4]. Partners previously uncovered many vulnerabilities as part of Project Glasswing, but achieving a sufficient reduction in cyber risk remains challenging because verifying, prioritizing, and fixing findings is complex [1]. Through the Cyber Mission, Anthropic deploys engineering talent and provides tools and funding for those who secure critical infrastructure and open-source software [1].

Anthropic also launched OSS Scanner to offer open-source projects regular security scans from its strongest models for free, targeting the shared code that most software depends on [3]. This tool focuses on finding vulnerabilities, proposing patches, and hardening underlying code [1]. The vendor strategy shifts focus from abstract alignment discussions toward practical deployment of frontier models directly inside industrial networks [1].

In our view, deploying software patches while plants keep running assumes industrial operators are entirely comfortable letting frontier models rewrite legacy code on the fly — a gamble that redefines operational risk.

Artificial IntelligenceCybersecurityAI ToolsAnthropic