AI vs. cyberattacks: CAREER award funds research to make cybersecurity smarter

When cybercriminals attacked Equifax in 2017, more than 140 million Americans had their financial data exposed. Incidents like that inspire Peng Gao, a security researcher and assistant professor of computer science, to build new cyberprotection systems. 

Gao and his team received a National Science Foundation Faculty Early Career Development Program (CAREER) award to explore how artificial intelligence (AI) can help stop future attacks.

The five-year, $625,000 grant will support Gao’s vision of intelligent, knowledge-enhanced, context-aware, and human-inspired cyberthreat protection powered by AI. The hope is that these tools can analyze complex threats faster and more accurately than human security professionals working alone can.

“Whenever there’s a new technology, both attackers and defenders adapt it,” Gao said. “Our mission is to make sure AI becomes a tool for protecting people, not just another way to attack them.”

Drowning in alerts

IT teams across every industry and government agency face a barrage of security alerts every day. Most are false alarms, but the work required to sort and categorize them can bury evidence of a serious breach until it’s too late to prevent harm. 

“You might get an alert today, but by the time you confirm it’s real, an attacker has already been inside your network for weeks or months,” Gao said.

Gao’s research will explore how large language models (LLMs) — the same technology behind OpenAI’s ChatGPT — could help security analysts identify and respond to threats more quickly and accurately. 

“It’s not about replacing people,” Gao said. “It’s about giving defenders a partner that can help them act faster and smarter.”

If successful, this approach could protect anyone who shops online, banks digitally, or entrusts data to a health care provider. Breaches at companies such as Walmart, Microsoft, or even hospitals put millions at risk of identity theft.

“Privacy is important to all of us,” Gao said. “If we can help companies keep customer data secure, that’s a win for everyone.”

Producing tools, not just papers

Gao has received seed grant support from the Commonwealth Cyber Initiative. And some of his earlier work has attracted attention from cybersecurity experts.

His team built a tool that converts cyberthreat intelligence reports into visual “cybersecurity knowledge graphs,” making it easier for threat analysts to identify attack patterns and create detection rules. The work was presented at the IEEE European Symposium on Security and Privacy earlier this year, and it’s being tested now in real-world enterprises. That’s just the start of the sweeping impact Gao is working toward.

“We’re not just publishing papers,” he said. “We’re developing practical methods and tools that defenders can actually use.”

It's the kind of work can accelerate global progress.

“Peng joins more than 20 of our faculty who have received this prestigious National Science Foundation award. Federal support through projects like Peng’s CAREER grant fuels the groundbreaking research in our department that drives global recognition and has a real positive impact on people’s lives,” said Christine Julien, head of the Department of Computer Science.

A growing threat

Smarter defenses can’t come soon enough. 

Cybercriminals are actively leveraging AI to blur the line between trustworthy technologies and deceptive tools used for fraud and theft. Phishing emails and malicious code that once required significant time and expertise can now be generated in seconds.

Cybersecurity remains an arms race. Every new technology creates opportunities for both defenders and for attackers.

“Security will never be finished,” Gao said. “But with the right use of AI, we can get closer to staying one step ahead.”