July 14, 2026
This interview was recorded at the AWS Public Sector Summit In Washington DC.
At one time, reacting to a cyber threat was enough; signatures and blocklists could cope. But AI is now amplifying every aspect of cyberwarfare, including speed and scope, leaving these traditional defenses in the dust. Today’s threats can be numerous and personalized, making old approaches worthless.
Patricia Titus, Chief Information Security Officer at Abnormal AI, discussed the challenges of transitioning from legacy to modern AI-driven security architectures at the AWS Public Sector Show.
During the interview, she advised focusing on asset visibility, decision-making platforms, and non-human identities to enhance cybersecurity.
So many federal organizations are deploying tools that it is difficult to get a baseline or inventory of what is on a system. Titus recommends starting with an accurate inventory to establish a “normal” baseline and detect abnormalities.
She also stressed the necessity of modernizing security architecture in parallel with cloud migration and the dangers of relying on outdated tools.
