White House launches AI Cyber Challenge to test how top AI models protect software

At the Black Hat USA conference in Las Vegas today — the nation’s largest hacking conference — the Biden-Harris administration announced the launch of a two-year open competition to explore how AI can be used to protect and defend the U.S.’s most vital software, including computer code that keeps the internet and critical infrastructure running.

The “AI Cyber Challenge” (AIxCC) is led by the Defense Advanced Research Projects Agency (DARPA), with collaboration from Anthropic, Google, Microsoft and OpenAI. The companies will offer expertise and make their models available for the challenge, which will feature nearly $20 million in prizes. To ensure broad participation and a level playing field, DARPA will also make $7 million available to small businesses that want to compete.

According to a press release, the AIxCC challenge “will demonstrate the potential benefits of AI to help secure software used across the internet and throughout society, from the electric grids that power America to the transportation systems that drive daily life.”

‘Clarion call’ to bolster security of critical software

In a media briefing, Arati Prabhakar, director of the White House Office of Science and Technology Policy, said, “This competition will be a clarion call for all kinds of creative people in organizations to bolster the security of critical software that American families and businesses and all of our society relies on.”

DARPA, he pointed out, has been issuing these kinds of prize challenges for nearly two decades to attract a “wide raft of talented people in organizations to compete to solve staggeringly hard problems. This is one of the ways that public and private sectors work together to do big things to change how the future unfolds, and that’s why the White House asked DARPA to take on the critical topic of AI for cybersecurity.”

The Open Source Security Foundation (OpenSSF), a project of the Linux Foundation, will serve as a challenge advisor. It will also help ensure that the winning software code is put to use right away protecting America’s most vital software to help keep the American people safe.

AI’s promising potential to secure vital computer code

Perri Adams, program manager at DARPA’s Information Innovation Office, said that cyber-defenders are “tasked with protecting really a daunting maze of technology, and today, they don’t have the tools capable of security at this scale.” Despite these vulnerabilities, modern advances in AI, when used responsibly, “have remarkable potential for securing our code. We have to keep the offense one step ahead. And AI offers a very promising approach for that.”

The press release added that the announcement is “part of a broader commitment by the Biden-Harris administration to ensure that the power of AI is harnessed to address the nation’s great challenges, and that AI is developed safely and responsibly to protect Americans from harm and discrimination.”

Last month, the administration secured voluntary commitments from seven leading AI companies to manage the short- and long-term risks of AI models. Representatives from OpenAI, Amazon, Anthropic, Google, Inflection, Meta and Microsoft signed the commitments at the White House.

The commitments include ensuring products are safe before introducing them to the public — with internal and external security testing of AI systems before their release as well as information-sharing on managing AI risks.

In addition, the White House says it is developing an executive order and will pursue bipartisan legislation to help America lead the way in responsible AI innovation.