The US government is launching the AI Cyber Challenge, a competition offering $20 million in rewards for using artificial intelligence (AI) to bolster cybersecurity for critical national infrastructure.
Leading AI firms, including Anthropic, Google, Microsoft, and OpenAI, are partnering in this initiative and providing their technology for the contest.
The announcement came during the Black Hat USA hacking event in Las Vegas.
After a qualifying round in the spring, 20 top teams will progress to the DEF CON 2024 semifinals.
The best five teams will then each receive $2 million and move to the DEF CON 2025 finals, with the leading trio vying for additional rewards, including a $4 million top prize.
All solutions must be open-sourced, with the Linux Foundation's Open Source Security Foundation advising.
The Defense Advanced Research Projects Agency (DARPA) will also offer $1 million to seven SMEs to ensure diverse participation.
This initiative follows DARPA's 2014 Cyber Grand Challenge.
The overarching aim is leveraging AI's potential to enhance US cybersecurity infrastructure. DARPA's Perri Adams emphasised the importance of staying ahead in defence, highlighting the unique opportunity this challenge presents for cybersecurity and AI experts.