1 in 5 finance leaders fear deepfake impersonation

Encountering fraud that looks like the workings of an artificial intelligence tool? You’re not alone.

According to a new survey of 230 senior financial leaders conducted by Early Warning Services’ fraud prevention brand Certos, as many as 81% of respondents said their company had experienced an attempted fraud that involved, or appeared to involve, “AI-generated or AI-enhanced content or techniques” within the last year.

A similar share (79%) said they were very concerned that AI-enabled fraud could negatively affect trust in their organization.

The potential repercussions of AI-powered fraud, evidently, are sweeping. In response to an open-ended question about their biggest concerns about AI-enabled fraud or scams, just about one in five respondents (21%) cited the potential for executive deepfake impersonation. The most common response to that question, cited by 43% of respondents, was payment diversion, followed by detection gaps or control lags, named by 38% of respondents.

As one anonymous CFO of a Fortune 500 company put it: “Biggest concern is a convincing fake request slipping through and sending company money to the wrong place.”

Certos conducted the online survey Aug. 7-21 and received 79 responses from leaders at Fortune 500 companies, 127 from venture-backed companies and 24 from charitable organizations. The survey didn’t include responses from banks and credit unions, investment managers, financial services or fintech firms or those working on AI model development or AI services.

Detection gaps

Fraudsters who use AI may be likelier to slip through the cracks, too. Eighty-four percent of respondents said that AI-enabled fraud is much or somewhat more difficult to detect than traditional fraud. And, where most respondents (87%) believed they were very or somewhat prepared for the “sophistication of AI-enabled fraud and scams,” only half were confident they could identify such fraud or scam attempts “before money leaves the organization.”

Certos’ report comes at a time when businesses across many industries continue to loudly and enthusiastically embrace new AI tools in the name of productivity, even as purveyors of the technology themselves continue to issue dire warnings about the threats posed by their own products. In a leaked prospectus document obtained by Reuters, for instance, Anthropic officials explicitly said that advanced AI could lead to “catastrophic or existential risks to humanity.”

In an email, a Certos spokesperson acknowledged that “AI can make fraud easier to create, scale and disguise.”

“But the bigger takeaway from the research is that this is becoming a commercial banking issue, not simply a cybersecurity issue,” the spokesperson said. “More broadly, we do not view these findings as a reason to retreat from technology. They show that the threat is evolving quickly and that financial institutions need to continue evolving their defenses along with it.”

Executives are certainly not the only ones facing AI-related fraud, with recent research by Consumer Reports showing that nine in 10 Americans have encountered at least one type of cyberattack or digital scam. “Scams are felt by everyone and rapidly evolving thanks to AI,” said Craig Newmark, founder of Craigslist and the “Take 9” cybersecurity public awareness campaign, in that report.

In a world of compounding fraud, driven by AI or otherwise, whose job is it to stop it? The Certos report suggests that companies are increasingly leaning on their financial institutions for some help. Ninety-three percent of respondents said that AI-enabled fraud risk is changing how their organization thinks about its financial institution partners.

When asked to name their top three expectations of their financial institutions, 51% of respondents said they’re looking for “better detection of suspicious activity.” Early Warning Services is owned by some of the top banks in the U.S.

In a statement, Certos General Manager Ben Chance said that the solution to AI-enabled fraud rests on “having the inter-bank network intelligence to help determine whether the receiving account actually belongs to the person or business the sender thinks they’re paying before money leaves the bank.”

In an email, a Certos spokesperson pointed out that the survey shows “strong agreement that this problem cannot be solved by any one institution or sector alone.”

Response to fraud, the spokesperson said, has to “extend across the ecosystem.”

“Telecom providers, payment networks, marketplaces, law enforcement, regulators and policymakers all have roles to play,” the spokesperson said.