AI Agents Face 10 Student Teams in Cyber Defense Contest, Exposing Limits
Updated
Updated · The New York Times · May 12
AI Agents Face 10 Student Teams in Cyber Defense Contest, Exposing Limits
2 articles · Updated · The New York Times · May 12
One blue team in the National Collegiate Cyber Defense Competition was made up entirely of AI agents, joining 10 regional-winning student teams in a live defense test against professional hackers.
Two days of attacks from a seven-member red team targeted a computer network in San Antonio, with defenders losing points whenever attackers breached machines or stole data using custom malware.
The Las Vegas event showed AI can assist both offense and defense but still makes mistakes and could not match seasoned cybersecurity professionals or top computer science students.
Run by the University of Texas at San Antonio, the contest was designed to mirror real-world cyberwarfare as AI is expected to play a growing role in cybersecurity.
With AI attackers evolving faster than human defenders, is perfecting human-machine teaming our only viable defense?
AI automates cyber defense, yet the workforce gap widens. What skills will humans need when machines are the first responders?
AI safety features cripple its defensive use. How do we build an effective AI guardian that can't become a weapon?