August 10 – The Biden administration is collaborating with four leading artificial intelligence firms to launch a new cybersecurity competition targeted at defending the United States’ critical infrastructure.
The “AI Cyber Challenge,” unveiled Wednesday at the Black Hat USA conference in Las Vegas, will connect specialists with AI models developed by Anthropic, Google, Microsoft, and OpenAI to create systems to find and cure software flaws. The DARPA-hosted contest offers a $20 million prize pool for the best technologies that might be deployed to secure government infrastructure ranging from transportation to electricity grids.
Teams are planned to compete in a qualifying event next spring, with the winners going on to compete in a semifinal tournament at Def Con the following year. The top five teams will participate in the finals, which will take place at Def Con 2025. Winners will then be urged to open source their systems so that they “can be used by everyone from volunteer, open-source developers to commercial industry,” Adams said on Tuesday.
Adams and the other challenge coordinators will hold a keynote, virtual news conference, and panel later this week to provide more information.
- Advertisement -
Google, Microsoft, Nvidia, OpenAI, and Meta agreed in May to make their language models available for public examination “by thousands of community partners and AI experts” during this year’s Def Con, which begins later this week.
The new challenge follows a week of cybersecurity pronouncements from the White House. The Biden administration hosted a symposium on K-12 ransomware assaults on Tuesday, announcing a range of initiatives to combat the menace. The Federal Communications Commission is scheduled to begin a pilot programme to provide $200 million in subsidies to K-12 schools and libraries over the next three years to strengthen their cyber defences. Amazon Web Services has contributed $20 million to support a grant programme for schools and state education ministries.
