Fighting AI-enhanced fraud as an ecosystem

At a webinar hosted in association with Ecommpay, industry participants discussed how AI-enhanced fraud has shifted to exploiting human vulnerabilities and urged coordinated regulation and data sharing across the payments ecosystem. Panelists referenced Interpol’s estimate that AI-enabled scams are about 4.5 times more profitable than traditional methods.

Speakers described how fraudsters now combine social engineering, deepfakes and automation to target individuals rather than technical systems. The panel pointed out that psychological tactics amplified by AI can scale attacks and make them harder to detect for both consumers and firms, limiting the effectiveness of individual controls and standalone customer education.

The regulatory and operational picture was described as fragmented. Responsibility for fraud prevention is split across payments regulators, financial conduct authorities, law enforcement and data protection agencies, with no single body overseeing the full lifecycle of fraud risk. Participants said that partial remits create gaps that organised criminals can exploit.

Several legal conflicts were identified as barriers to industry collaboration. Data protection rules and privacy requirements can restrict the exchange of suspicious-activity information between companies. Competition and antitrust concerns may discourage banks and fintechs from pooling intelligence. Differences in cross-border reporting standards and supervisory remits further complicate rapid information sharing.

Panelists outlined specific reforms that could address those gaps. Suggestions included creating legal safe harbours or sector-specific exceptions for sharing threat indicators, harmonising reporting standards to reduce duplicate compliance burdens, and establishing neutral, non-profit hubs to collect and redistribute anonymised intelligence while preserving privacy safeguards.

The idea of standardised fraud processes received mixed reactions. Advocates said common procedures could speed incident reporting and pattern recognition across firms. Critics warned that a single standard could impose heavy costs on smaller businesses and may not adapt quickly enough to new AI-driven attack methods.

Operational and commercial obstacles were also discussed. Legacy IT systems, a lack of shared taxonomies for fraud data, potential reputational harm from intelligence sharing, trust deficits between competitors and the cost of building shared platforms were all listed as practical roadblocks.

Panelists recommended proportionate measures, including tiered reporting requirements, funding or subsidies to help smaller firms join shared intelligence systems, independent governance for data hubs and international cooperation to align cross-border rules. Contributors to the webinar included Willem Wellinghoff, UK Chair and Chief Compliance Officer at Ecommpay, with Teresa Connors as moderator.

Interpol’s assessment of rapidly evolving financial fraud was cited as the backdrop for the discussion. Participants concluded that changes in law, shared technical standards and trusted data-sharing mechanisms would be needed for firms and regulators to coordinate responses to AI-enabled scams.