AI/ML’s Role in Cybersecurity: Balancing Innovation, Safety

AI and machine learning systems are reshaping cybersecurity by addressing complex challenges that traditional methods struggle to solve, according to Trail of Bits’ Michael Brown.

See Also: NHS Ransomware Attack: Healthcare Industry Infrastructures Are Critical

Using AI and ML systems on their own or in conjunction with conventional systems allows organizations to overcome the limitation of Moore’s Law and make new breakthroughs and advances that weren’t possible before, Brown said. But these advancements come with their own set of security concerns, and Brown said rigorous security policies are needed from the data collection phase to stop vulnerabilities.

“The AI/ML systems that we build for achieving security objectives – they tend to excel where our conventional techniques start to fail, and they start to fail where the conventional techniques tend to succeed,” Brown said. “Complementary approaches that use conventional techniques – where they’re strong – and AI/ML techniques – where they’re strong – get a ‘best of both worlds’ approach.”

In this video interview with Information Security Media Group at DEF CON 2024, Brown also discussed:

• How AI/ML systems can complement conventional security methods;
• Unique security challenges posed by AI/ML systems, particularly in LLMs;
• Current efforts in government and the private sector to secure AI and ML technology.

Brown works on research projects focused on security-oriented software analysis and transformation. His primary research interest is the development of software transformation techniques to improve the security of computing systems. Prior to his work in software security, Brown spent eight years in the U.S. Army, where he served as a UH-60M pilot and aviation mission survivability officer.