Combating Fraud and Scams in the Age of AI

The Problem

“Fraud and scams are not only a threat to the European economy, they are also a significant cause of financial and psychological distress for European consumers.”

– Commissioner Michael McGrath[1]

Financial fraud and scam in Europe have reached unprecedented sophistication, shifting from pure payments abuse to human-centered manipulation amplified by generative AI. For example, total payment fraud losses across the EU/EEA rose to €4.2 billion in 2024 from €3.5 billion in 2023[2].
ENISA and Europol point to a growing use of AI in fraud and scams, as well as the increasing industrialization of fraud, which is increasingly offered as a service across borders.^[3] A report by identity platform Signicat points out that in 2024 already 42.5% of all detected fraud attempts in the financial and payments sector were AI driven.^[4]

Social engineering, phishing, and remote card and credit transfer fraud now dominate, with AI enabling highly convincing scams through voice cloning, deepfakes, and hyper-personalized lures. These attacks exploit human factors such as urgency and hijacked trust, as well as weak verification and synthetic identities. At the same time,  attacks are enabled and scaled by systemic factors such as Crime-as-a-Service models and cross-border mule networks. Furthermore, social engineering scams bypass strong customer authentication (SCA) and traditional controls, and instant payments create narrow windows of opportunity for fraud prevention.  Fraud is increasingly a connected, systemic problem rather than a set of isolated cases, making it harder for banks, telecoms, governments, and law enforcement to respond in a coordinated way.

The Challenge

Design a solution that helps to protect people and organizations from AI-powered financial fraud and scams.

Things to consider

• Think about various types of fraud and scams – across cards, A2A credit transfers (including instant), e-money wallets, and adjacent journeys like onboarding, account recovery, disputes, customer communications, and society at large.
• Consider the specific needs of your target audience, which may be variouspersonas such as victims, fraud analysts, call-center agents, or regulators. Also, your solution can be aimed at both private and/or public-sector bodies, including banks, payment service providers (PSPs), regulators, law enforcement, and other relevant stakeholders.
• What types of fraud or scam your solution address and what indicators would show that it is working: e.g., reducing losses, detecting fraud earlier, improving accuracy, saving time for teams, stopping scam interactions and risky transactions.
• Note that currently manipulation outpaces controls. For example,  strong customer authentication helps, but fraudsters still convince victims to authenticate. Controls must recognize human risk in all elements of financial services.
• Instant payments make controls even more challenging. Consider solutions that stop scams before money is transferred.
• Cross-border complexity: Fraud complexity increases when counterparties are outside the EEA. The design should consider counterparty risk and geographies.

[1] https://commission.europa.eu/news-and-media/news/commission-gathers-key-european-enforcement-network-and-stakeholders-fight-online-consumer-fraud-2025-02-21_en

[2] https://www.ecb.europa.eu/press/pr/date/2025/html/ecb.pr251215~e133d9d683.en.html

[3] https://www.enisa.europa.eu/publications/enisa-threat-landscape-2025 https://www.europol.europa.eu/cms/sites/default/files/documents/Internet%20Organised%20Crime%20Threat%20Assessment%20IOCTA%202024.pdf

[4]https://www.signicat.com/press-releases/42-5-of-fraud-attempts-are-now-ai-driven-financial-institutions-rushing-to-strengthen-defences?utm_source=chatgpt.com