AI fraud race may strain financial security, Plaid CEO says

What happened

Zach Perret, chief executive of the data aggregator Plaid, has warned that the spread of artificial intelligence could make the coming year a difficult one for financial security. Speaking at the Money20/20 conference this week, he said the industry is in an intense race to use AI to counter fraud that is itself increasingly AI-driven.

Banking Dive reported the remarks, and its coverage framed the issue around a headline concern: that AI's growth could make the next year a "scary" one for people's financial security. Perret's comments were the centrepiece of that reporting.

The available reporting is limited to this single source. It does not detail the specific fraud patterns, data or countermeasures Perret referred to beyond his core remarks, so the picture here is necessarily partial.

Why it matters

The point at the heart of the story is that AI now sits on both sides of the fraud contest. The same capabilities that help banks and fintechs detect suspicious activity are available to those trying to get around those checks. For organisations running digital financial services, this means defence is no longer a static rulebook but a continuous, fast-moving operational task.

For leaders in experience and digital transformation, the implication is that security and customer experience are now tightly coupled. Every additional verification step, alert or declined transaction is a moment of friction for a customer, and the volume of those moments will be shaped by how well AI-driven defences are designed and tuned.

The Renascence take

Most readers will hear this as a security story about attackers. The more useful reading is that fraud is a trust problem expressed through experience. Customers rarely judge their bank on how sophisticated its models are; they judge it on whether a blocked payment is explained clearly, whether a genuine transaction gets through, and whether they feel protected rather than policed.

The behavioural principle underneath is that people tolerate friction when they understand its purpose and trust its accuracy, and they abandon services when friction feels arbitrary. A customer-obsessed operator should therefore measure false positives and recovery journeys as seriously as fraud losses. It should also make every security intervention explainable in plain language, so that protection reads as care rather than suspicion.