Four in Five Finance Leaders Report AI Fraud Attempts as Deepfake Fears Grow

A convincing fake payment request could be enough to send company money to the wrong account. Four in five surveyed finance leaders said their organisations had faced attempted fraud involving, or apparently involving, artificial intelligence in the previous 12 months, highlighting concerns about the growing sophistication of corporate scams.

The survey, published on 8 October 2026 by Certos, the fraud prevention brand of Early Warning Services, found that 81% of respondents reported such attempts. Meanwhile, 79% were very concerned that AI-enabled fraud could damage trust in their organisations.

Payment diversion was the most frequently cited concern, while executive deepfake impersonation featured in responses from 21% of leaders. The survey did not establish whether concerns about deepfakes had increased over time.

Payment Diversion Tops Corporate Fraud Concerns

Payment diversion was cited by 43% of respondents asked to describe their biggest concerns about AI-enabled fraud affecting corporate finance, payments, and banking relationships. Detection gaps or delays in existing controls ranked second, at 38%, followed by phishing and social engineering at 31%. Executive deepfake impersonation was cited by 21% of respondents.

Other concerns included vendor or bank impersonation, cited by 14%, and fake invoices or documents, mentioned by 13%. Data and privacy breaches accounted for 8%, while 7% highlighted potential damage to trust and reputation. One anonymous chief financial officer at a Fortune 500 company described the risk as a convincing fake request slipping through checks and sending company money to the wrong place.

These percentages reflect respondents' stated concerns, not confirmed successful fraud cases or verified financial losses.

Deepfake Fears Put Executive Impersonation Under Scrutiny

AI-generated audio, video, and written communications can be used to imitate people whom employees recognise and trust. In corporate finance, a fabricated instruction that appears to come from a senior executive could pressure an employee to authorise a fraudulent transfer.

The finding that 21% of respondents cited executive deepfake impersonation highlights concern about fraudulent instructions appearing to come from trusted leaders. However, the survey did not establish how many respondents had experienced a successful deepfake attack.

Payment diversion ranked higher overall, indicating that leaders were also concerned about the financial consequences of fraudulent transactions, regardless of the technique used.

Most Leaders Say AI Fraud Is Harder to Detect

The survey found that 84% of respondents considered AI-enabled fraud somewhat or much harder to detect than traditional fraud. At the same time, 87% said their organisations were somewhat or very prepared for the sophistication of AI-enabled fraud and scams. Yet only half felt confident they could identify a fraudulent attempt before money left their organisation.

The findings point to a gap between perceived preparedness and confidence in identifying an attack before a payment is made. A Certos spokesperson said AI could make fraud easier to create, scale, and disguise. The spokesperson described the issue as a commercial banking concern rather than simply a cybersecurity problem. The spokesperson added that the findings were not a reason for businesses to retreat from technology, but showed the need for financial institutions to keep adapting their defences.

Businesses Expect Banks to Strengthen Fraud Detection

The survey found that 93% of respondents said AI-enabled fraud risk was changing how their organisations viewed their financial institution partners. When asked about their main expectations of those partners, 51% called for better detection of suspicious activity. Ben Chance, Certos general manager, said cooperation between banks could help establish whether a receiving account belonged to the person or business a sender intended to pay before funds left the bank.

A Certos spokesperson said the problem could not be solved by any single institution or sector. The organisation identified telecoms providers, payment networks, online marketplaces, law enforcement agencies, regulators, and policymakers as having roles to play. The survey was conducted online between 7 and 21 August 2026. It included 79 respondents from Fortune 500 companies, 127 from venture-backed businesses, and 24 from charitable organisations.

It excluded respondents working for banks and credit unions, investment managers, financial services, and fintech firms, as well as those involved in AI model development or AI services. The results therefore reflect the views of the surveyed corporate finance leaders, not the entire financial sector.

AI Fraud Creates a Wider Challenge for Businesses

For corporate finance teams, the central challenge is to identify fraudulent instructions before transferring money. The findings highlight the importance of fraud detection, payment verification and cooperation between financial institutions and other organisations involved in transactions. The survey does not prove that AI was responsible for every suspected incident. However, the results show that many surveyed finance leaders are concerned about the potential for AI-enabled techniques to make fraud harder to detect.