![[2602.14783] What hackers talk about when they talk about AI: Early-stage diffusion of a cybercrime innovation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.14783] What hackers talk about when they talk about AI: Early-stage diffusion of a cybercrime innovation
Summary
This article explores how cybercriminals are discussing and utilizing artificial intelligence (AI) to enhance their operations, revealing insights into the early-stage diffusion of AI in cybercrime.
Why It Matters
As AI technology rapidly evolves, understanding its implications for cybercrime is crucial for law enforcement and policymakers. This research highlights the dual nature of AI as both a tool for criminals and a challenge for security, emphasizing the need for proactive measures to mitigate risks associated with AI in cybercrime.
Key Takeaways
- Cybercriminals are increasingly curious about AI's potential for illicit activities.
- Discussions among hackers reveal both enthusiasm and skepticism regarding AI's effectiveness in cybercrime.
- The study provides insights into how AI tools are being misused and tailored for criminal purposes.
- Understanding these trends is vital for developing effective countermeasures by law enforcement.
- The research employs a unique dataset from cybercrime forums to analyze conversations over seven months.
Computer Science > Computers and Society arXiv:2602.14783 (cs) [Submitted on 16 Feb 2026] Title:What hackers talk about when they talk about AI: Early-stage diffusion of a cybercrime innovation Authors:Benoît Dupont, Chad Whelan, Serge-Olivier Paquette View a PDF of the paper titled What hackers talk about when they talk about AI: Early-stage diffusion of a cybercrime innovation, by Beno\^it Dupont and 1 other authors View PDF Abstract:The rapid expansion of artificial intelligence (AI) is raising concerns about its potential to transform cybercrime. Beyond empowering novice offenders, AI stands to intensify the scale and sophistication of attacks by seasoned cybercriminals. This paper examines the evolving relationship between cybercriminals and AI using a unique dataset from a cyber threat intelligence platform. Analyzing more than 160 cybercrime forum conversations collected over seven months, our research reveals how cybercriminals understand AI and discuss how they can exploit its capabilities. Their exchanges reflect growing curiosity about AI's criminal applications through legal tools and dedicated criminal tools, but also doubts and anxieties about AI's effectiveness and its effects on their business models and operational security. The study documents attempts to misuse legitimate AI tools and develop bespoke models tailored for illicit purposes. Combining the diffusion of innovation framework with thematic analysis, the paper provides an in-depth view of emergin...