Anthropic launches defense program for critical infrastructure

Anthropic has introduced a cybersecurity initiative called the Cyber Mission aimed at protecting essential public systems and shared software code. As part of this push, the artificial intelligence developer is launching a dedicated program for critical infrastructure alongside a free scanning service for open source software. The company announced that it will provide advanced artificial intelligence models, specialized security research, and technical staff to major cybersecurity firms and industrial manufacturers.

Most utility networks, including electrical grids, municipal water services, and public transport systems, rely on aging computer networks known as operational technology. Because these systems often run continuously and cannot easily be turned off for maintenance, known security flaws frequently remain unpatched for years. If hostile groups compromise these networks, everyday essentials like power delivery or clean drinking water could face severe disruptions. By giving defenders better tools to spot weaknesses early, Anthropic hopes to prevent real world outages before bad actors take advantage of them.

Despite these efforts, deploying artificial intelligence to defend physical utilities carries inherent risks. Finding software vulnerabilities is only part of the problem, as applying fixes to delicate industrial machinery without causing accidental downtime remains extremely tricky. Furthermore, because the same artificial intelligence capabilities can potentially be exploited by foreign adversaries to automate attacks, it remains an open question whether software automated defenses can truly outpace sophisticated cyber threats in the long run.