Written by Vicomtech
In the ever-evolving digital landscape, fraud has become not only more frequent but also significantly more sophisticated. With the rise of artificial intelligence (AI), particularly large language models (LLMs), criminals are leveraging advanced tools to deceive victims more convincingly than ever. As a result, ongoing research into fraud detection is now more crucial than ever to protect individuals, businesses, and public institutions across Europe.
According to Europol’s Internet Organised Crime Threat Assessment (IOCTA) 2025, online fraud is poised to become the dominant form of serious and organised cybercrime in the coming years. Europol warns: “Europol expects online fraud to outpace other types of serious and organised crime as it is being accelerated by AI, aiding social engineering and access to data.”
LLMs are enabling attackers to craft highly realistic phishing emails, impersonate individuals in real-time through chat or voice interfaces, and even generate fake yet credible documentation. These tools lower the technical barriers for fraud and expand the scale at which scams can be launched. Combined with access to stolen personal data and weak cybersecurity awareness, the result is a perfect storm for cyber-enabled fraud.
The IOCTA 2025 highlights how fraud is no longer an isolated issue; it is now connected to other cyber threats such as ransomware, cryptocurrency abuse, and digital identity theft. The report notes a rise in crime-as-a-service ecosystems, where fraud tactics, stolen credentials, and even AI-generated content are bought and sold on underground markets.
In this context, research is essential, not only for developing detection tools but also for understanding attacker behavior and anticipating new trends. As cybercriminals become more innovative, so too must our defenses. The insights from IOCTA 2025 underscore the urgency of accelerating research and cooperation to stay ahead of the threat and to ensure a safer digital future for all.
