While bad actors are creating synthetic identities to commit fraud and launder money, FIs can use synthetic identities that they control to test their systems, making the FIs less vulnerable to genuine attacks.

In identity fraud, information is stolen to open new accounts or lines of credit in a real person’s name. Alternatively, a synthetic identity is a digital persona built from a mix of real and fake data, often used by bad actors to bypass security checks at financial institutions.

“The [synthetic] IDs are getting so, so good … because of the ability of gen AI and LLMs to generate fake biometric data paired with fake documents,” Jose Caldera, chief executive of tech firm Yanez Compliance, told FinAi News.

“Every so often there is something that fundamentally can change the way you think about security, and I think that AI in general is fundamentally questioning the use of biometric data for the purposes of security,” he said.

Unsecured credit synthetic identity fraud losses reached about $2.9 billion in the United States in 2025, up from $1.8 billion in 2020, according to a June 10 report released by fraud-prevention tech company Mitek Systems and financial advisory firm Datos Insights. That figure is expected to rise to $3.1 billion for 2026.

AI is a big reason for the increase in synthetic identity fraud.

“Limitations of scale [for threat actors] are quickly disappearing, because you’re no longer constrained by how many people can join in on your fraud scheme,” Madhu Nadig, co-founder and chief technology officer at tech firm Flagright, told FinAi News. “Now you can just spin up some synthetic identities.”

Fighting fire with fire

Yanez uses AI-generated synthetic identities to help FIs test and validate identity verification systems. The Yanez Compliance platform can create localized identities with characteristics from Latin America, Asia, Africa or any location of the FI using the platform, Caldera said. Yanez has “10s of customers” but Caldera declined to name them. Yanez began selling its product at the end of 2024.

The company uses synthetic identities that produce adversarial data — false data created to trick or manipulate an AI system. Yanez’s adversarial data exposes system vulnerabilities in a controlled environment, Caldera said.

“Organizations are going to have to deal with this because identities are becoming so easy to fake,” Caldera said. “The risk to financial institutions and identification providers is in not using synthetic identities.”

When building identity verification systems in the past, Caldera and his colleagues focused on “positive identification,” since 99% of identity verification cases tend to be real people conducting legitimate transactions, he said.

“But we never spent that much time in those edges and in the 1%, [and] that’s what adversarial data is about,” Caldera said. “It’s about how you test those edges.”

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas. This inaugural event will include speakers from Fifth Third and Capital One as well as a fireside chat with Piermont Bank founder and Chief Executive Wendy Cai-Lee.