Industry calls for unified approach to AI-driven fraud

At a webinar hosted in association with payments firm Ecommpay, industry experts described a widening gap between the speed and scale of fraud and the fragmented defences of banks, merchants and regulators. The session included Willem Wellinghoff, UK chair and chief compliance officer at Ecommpay, with Teresa Connors moderating.

Panelists reported that fraud has shifted from purely technical attacks to schemes that exploit human psychology, using social manipulation to bypass controls. They said generative AI and other automated tools amplify criminals’ reach and efficiency. The panel noted research cited by Interpol that finds AI-enhanced fraud can be 4.5 times more profitable than traditional methods and quoted Interpol’s assessment that financial fraud is ‘one of the world’s most severe and rapidly evolving transnational crimes, with significant economic and human consequences.’

Speakers identified two structural weaknesses. First, regulatory responsibilities for fraud prevention are split across multiple bodies, leaving no single entity with end-to-end oversight. Second, firms have built different protection tools and processes but share limited data, reducing the ability to detect organised schemes that cross institutions and channels.

To address those gaps, panelists proposed regulatory reform that sets common outcomes and reporting standards while scaling obligations to reflect firm size and transaction volume. They described a risk-based framework intended to harmonise expectations across financial regulators, consumer protection agencies and law enforcement without imposing identical requirements on all businesses.

On operational fixes, experts outlined practical steps for industry collaboration. Proposals included common taxonomies for fraud types, standard incident-reporting formats and shared non-profit platforms to aggregate anonymised signals and indicators. Such measures aim to help institutions identify cross-party patterns more quickly and avoid duplicated effort.

Panelists also listed legal and operational barriers. Data protection rules and competition law can limit information exchange, and differing cross-border jurisdictions complicate international cooperation on payments. Technical interoperability, integration costs for smaller firms and concerns about reputational or regulatory exposure may discourage firms from sharing breach information. The group said any standard process would need safeguards for privacy and proportionality.

As an alternative to competing proprietary systems, the panel suggested exploring a neutral, non-profit body to coordinate shared services and host anonymised intelligence. Speakers noted governance, funding and legal liability would be significant issues to resolve before such an entity could operate broadly.

Discussion of synthetic content and automated tooling emphasised that scams are becoming harder for both consumers and detection systems to distinguish from legitimate activity. Panelists outlined a need for cross-sector collaboration, clearer regulatory roles and safe channels for sharing fraud signals to adapt to the changing threat environment.