Nat Sec Press Updated on May 12, 2026 2 min read

Episode 002: Crimes of Influence

In this episode of the Nat Sec Press Podcast, Dr Nicole Matejic unpacks the paper Crimes of Influence: Generative Artificial Intelligence-led Crime as a Service (CaaS), exploring how generative AI is changing cybercrime.

What if the future of cybercrime isn’t just about exploiting systems — but exploiting human decision-making itself?

Drawing on behavioural economics, psychology and cybercrime research, this episode explores concepts including:

• Social engineering
• Crime-as-a-Service (CaaS)
• System 1 and System 2 thinking
• Heuristics, biases and decision-making shortcuts
• Persuasive AI and influence at scale; and
• Why generative AI changes the sophistication and precision of cyber-crime

The episode also examines how generative AI is making malicious influence more personalised, convincing and difficult to detect - exploiting the way humans naturally process information, trust others and make decisions.

This paper was co-authored with Chris Wilson, from the University of Auckland.

This podcast is part of the Nat Sec Press Podcast series, which breaks down academic research, books, and national security issues into concise, plain-English discussions.

Links:

Crimes of Influence: Generative Artificial Intelligence-led Crime as a Service

‘Crimes of Influence’ – crimes that seek to influence people towards harmful outcomes – will be one of the defining features of generative artificial intelligence (AI)-led cybercrime. With an ability to persuade and influence at potentially unavoidable economies of scale, crimes of influence leverage the heuristics and biases that form part of everyday human cognition in ways that mislead, deceive, impair, disrupt, degrade and/or deny user-normative decision-making.