Dr Celia Chen - University of Maryland

The rapid introduction of AI-driven tools into cybersecurity work has changed how professionals in the field do their jobs. Generative AI systems are unavoidable in the information workspace, increasingly embedded in both commercial and in-house tooling that security teams rely on every day. Expertise in cybersecurity depends on intuition built through repeated hands-on engagement with difficult problems. While questions of user trust in these systems have received considerable attention, very little is yet known about the long-term consequences of systems operators relying on this type of automation software. If AI systems absorb the routine work with editing tools that traditionally trained this judgement, cybersecurity as a profession will need to develop its expertise differently, and the intelligence community will need to understand how this training can effectively take place. 

Dr Chen's research examines how the use of AI-driven support systems affects the development of intuition and subject matter expertise in cybersecurity professionals. The study draws on in-depth interviews with practitioners across a broad cross-section of the field, including US national laboratories, major technology companies, the financial sector, and critical infrastructure protection. The interviews explore which tools professionals actually use, how these tools have changed the nature of their daily work, when practitioners trust or override machine recommendations, and how they maintain and develop their skills as their roles shift from performing analysis to overseeing automated output. 

The findings will help the intelligence community anticipate how workforce expertise may change as AI adoption accelerates across its analytic tooling. The retention of skills and abilities in cybersecurity analysis is key to the continued effective function of security organisations as systems evolve over time, and the threat of advanced automated attacks escalates. This research will help to inform the design of AI-driven support systems that strengthen rather than erode human judgement, contributing to the broader question of how to gain the benefits of machine assistance while preserving human expertise these systems cannot replace.