DARPA Wants to Develop AI That Can Patch Security Vulnerabilities in Critical Software


 The world is now run on computers with thousands of software needed to operate elevators, lights, security systems and telecommunications. Lately it is often heard that personal data and important systems such as gas pipes are hacked causing disruption to daily activities.



Recognizing that reliance on this software exposes the United States to hacker attacks, DARPA announced the AI Cyber Challenge (AIxCC) to develop artificial intelligence (AI) that can detect before patching security vulnerabilities in critical software.


The competition will run for two years with Anthropic, Google, Microsoft, and OpenAI will provide infrastructure assistance to AIxCC participants to develop AI for cybersecurity purposes.



The Open Source Security Foundation (OpenSSF) will be an advisor to participants to create AI systems capable of dealing with important cyber security issues, such as the security of the United States' critical infrastructure and software supply chains.


The competition was announced weeks after U.S. intelligence officials alleged Chinese-made malware was hidden in networks that control power grids, communications systems and the water supply of U.S. military bases around the world. In a war situation, it can be activated to affect the availability of troops to oppose the enemy.

Previous Post Next Post

Contact Form