The increasing use of AI by romance fraudsters is making scams more efficient and harder to detect, according to new research by the Centre for Emerging Technology and Security (CETaS) at The Alan Turing Institute.
The scope and cost to victims of romance fraud has surged in recent years as AI has introduced new efficiencies that enable fraudsters to easily create more convincing personas, automate large-scale outreach and refine psychological manipulation.
Generative AI enables the mass production of synthetic personas with fake images and fabricated social histories that can bypass traditional detection methods, such as reverse image searches. AI-driven psychological profiling is also allowing fraudsters to identify victims and target their vulnerabilities more efficiently.
The research found that AI is being integrated into romance fraud in multiple ways, including the use of deepfakes to generate personalised dating profiles, and AI-generated audio and video content to enhance deception. Large language models (LLMs) are also assisting in improving the scripts used to deceive victims, and AI translation tools allow scammers to engage victims in multiple languages.
Although human scammers are still required to guide AI-generated outputs and correct their inconsistencies, AI-generated introductions significantly reduce the manual effort required to target potential victims.
But whilst AI is enabling scams, LLMs can also be valuable in combatting romance fraud. Their ability to create convincing scams suggests that they could in the future detect scam messaging.
As with other examples of online criminality, techniques are developing at a pace which bypass current defences as explored in a recent paper by CETaS. The paper forms part of a broader CETaS research project examining the role of AI in online criminality.
Lead author Simon Moseley, CETaS Visiting Research Fellow and Principal Data Scientist at the Home Office, said: “The use of AI in romance scams is developing quickly, helping criminals integrate new technologies into existing fraud networks with recent cases extracting millions from those seeking genuine connection. Often vulnerable victims are coming to psychological as well as financial harm through AI-assisted fraud. As romance scams grow in scale and sophistication, there has to be a greater focus on better detection capabilities and stronger safeguards.”
You can read the full briefing paper here.
