The UK National Cyber Security Centre (NCSC) published a new chapter in its research problem book on October 20, 2026, focusing on the security challenges and opportunities created by artificial intelligence.
Focusing AI cyber research
The NCSC announced that its research problem book now includes a chapter titled 'AI in cyber security problems'. Written by Paul J, Technical Director for Cyber AI Research at the NCSC, and Brad A, Principal Data Science & AI Researcher, the update addresses how frontier AI capabilities affect both cyber defenders and attackers.
According to the authors, traditional cyber security practices such as vulnerability management, incident response, and risk management were developed before modern AI existed and must evolve to account for AI-enabled systems and adversaries.
Three core research themes
The new chapter identifies seven specific research areas across three overarching themes. These include using AI to strengthen cyber defense, securing and assuring AI systems themselves, and leveraging data-driven methods to improve cyber resilience at scale.
The publication notes that newer AI technologies, specifically generative AI and agentic AI, differ from earlier machine learning tools because they can plan, reason, use tools, and take actions with increasing autonomy.
Human oversight and feedback
A central element across all identified research areas is human-AI collaboration. The NCSC highlighted the need to determine where autonomy provides value, where human judgment remains essential, and how oversight can be maintained during machine-speed operations.
The agency is seeking feedback on the identified challenges, suggestions for additional research questions, and examples of planned or ongoing research from government, industry, and academia.
Key facts and where they come from
- The NCSC expanded its research problem book with a new dedicated AI chapter.
we’ve re-written our NCSC research problem book to include an entirely new chapter 'AI in cyber security problems’.
- The AI chapter outlines seven areas covering three main themes.
Our new problem book chapter on ‘AI in cyber security’ identifies seven areas spanning three broad themes:
- Generative and agentic AI systems possess autonomous capabilities such as planning and tool use.
The new generation of AI systems (particularly generative AI and agentic AI) can plan, reason, use tools, and take actions with increasing autonomy.
