A team led by Texas A&M University was one of seven teams selected to head to the finals of the Defense Advanced Research Projects Agency’s (DARPA) Artificial Intelligence Cyber Challenge (AIxCC) — a two-year competition that brings together the best and brightest from AI and cybersecurity. The finals will take place in August of 2025 at DEF CON.
“AIxCC provides students with a unique opportunity to develop advanced technical skills, grow professionally and make meaningful contributions to pressing societal challenges,” said Dr. Jeff Huang, the leader of Texas A&M’s team and an associate professor in the computer science and engineering department.
The competition focuses on enhancing cybersecurity efforts to safeguard software critical to all Americans by developing protect software, finding and fixing vulnerabilities in AI design, and finding data protection solutions. It encourages students to use AI tools and push the boundaries of what AI can achieve, while positioning themselves as innovators who will help define the next generation of technological breakthroughs.
Each team was responsible for developing a CRS. To test the system, DARPA created challenge projects containing common and uncommon software vulnerabilities. DARPA then graded the CRS on the variety and accuracy of software patches it deployed, as well as the speed and quality of the repair.
Huang’s team developed a fully autonomous Cyber Reasoning System (CRS) using large language models (LLMs), securing a top seven spot in the semifinal against 40 teams. The CRS automatically detects and patches security flaws in complex systems like Nginx, the world’s leading web server, without human intervention.
For their success, the team will receive $2 million and have one year to further develop their technology.
“Competitions like AIxCC simulate real-world pressure, where quick and accurate responses are critical to success,” said Huang. “Working under such conditions helps students build resilience and adaptability, qualities that will serve them well in future careers, especially cybersecurity, where stakes are high.”
Led by Huang, the team includes Ze Sheng and Neha Manghnani, students at Texas A&M, Dr. Heqing Huang from City University of Hong Kong and an industry expert.
“The AIxCC competition challenged us to tackle some of the most pressing problems in cybersecurity,” said Sheng. “DARPA provided a range of complex tasks, allowing us to explore the frontiers of this field. I feel privileged to be part of the ‘all_you_need_is_a_fuzz_brain’ team and have had a chance to work with leading researchers in the field of cybersecurity to solve these problems.”
Moving forward, Huang believes this competition is critical as AI advances.
“As AI continues to evolve, it will drive advancements in areas such as healthcare, finance, and beyond,” said Huang. “Students who learn to use AI effectively will be better prepared to adapt and thrive in this rapidly changing landscape, making them highly valuable in the workforce.”