Persona, a global identity verification and fraud defense platform, released new data showing that identity fraud is becoming more adaptive as fraudsters use AI and automation to make repeated verification attempts at high speed.

High-velocity attacks, defined by Persona as five or more attempts within two minutes, are already common across certain types of cyberattacks and ecommerce fraud. Persona’s data shows this pattern is now playing out in identity, with bad actors using fake and stolen identities to make repeated, high-velocity verification attempts. Across Persona customers experiencing these attacks, from Fortune 500 companies to leading AI labs and large consumer tech companies, the data shows:

• The number of devices associated with high-velocity identity verification attempts more than doubled year over year. Often, identity thieves make minor changes before resubmitting an application after an attempt fails.
• Attackers’ devices were associated with 1.5x as many IP addresses as legitimate users. Fraudsters commonly “rotate” IP addresses using proxy networks to disguise their location and make it look like they’re using a new device.
• An 8x increase in devices that cycled through 10 or more IP addresses. The more sophisticated attacks often use residential proxies, which may belong to people who were tricked into installing malware.

Marketing Technology News: MarTech Interview with Mark Listes, CEO @ Pendulum Intelligence

The shift comes as AI and low-cost digital tools make large-scale attacks easier to automate. INTERPOL’s 2026 Global Financial Fraud Threat Assessment warns that these technologies are lowering barriers to sophisticated fraud, while Juniper Research forecasts that global fraud losses for financial institutions will increase 153% between 2025 and 2030, from $23 billion to $58.3 billion.

“Attackers are getting better at making repeated attempts look unrelated to one another, which lowers the difficulty for them to easily reach your users,” said Daniel Lee, Product Lead at Persona. “The challenge for fraud teams defending their users is understanding the activity associated with identity thieves: how quickly attempts are occurring, where they’re coming from, what infrastructure they’re cycling through, and whether seemingly separate interactions are actually connected. As attacks become easier to automate and faster to adapt, being able to connect those signals becomes increasingly important to stopping fraud at scale.”

Faster iteration by attackers also puts new pressure on fraud prevention teams. The traditional approach of taking days or weeks to observe a new attack pattern, determine which signals matter, test a response, and monitor the deployment is often too slow. For example, one Persona customer experienced repeated bursts of more than 150 account-creation attempts in 10-minute periods. The fradusters changed their methods every two to three hours and kept up the attack at all hours throughout the week in an attempt to get past Persona’s defenses.

Marketing Technology News: How MarTech Is Enabling Autonomous Brand Engagement Across Channels?

Persona is releasing new and updated fraud prevention products to help defend against these increasingly common attack vectors:

• Rapidly shifting, high-velocity attacks: Fraudsters cycle through devices, networks, and identities during high-velocity attacks. They also probe defenses at different companies and share tactics with other fraudsters. Many organizations work with Persona’s experts to configure the platform and create different user flows. Persona’s new Fraud Fight Flow is a fully managed user flow for customers that Persona continually updates based on its cross-industry threat analysis.
• AI-powered identity theft and fraud: Fraudsters are using AI in many parts of these attacks, including phishing schemes to steal victims’ identities and selfies. They can even turn the portrait from a stolen ID into an AI avatar and then submit a “selfie” as the victim. Persona’s research-backed risk signals are highly effective at flagging fraudulent activity and content. Organizations can also use Graph, a link analysis tool, to uncover connections between similar devices, users, and content. It’s especially helpful at detecting large fraud rings that submit AI-generated content and pay or coerce people into taking selfies.
• Activity from stolen accounts and identities: There are active markets for stolen identities and accounts that aren’t associated with risky activity. Persona’s Sentinel helps organizations assess risk before, during, and after users create an account or log in. They can use these insights to help stop account takeovers and determine when to add additional verification requests.

As AI makes attacks easier to automate and adapt, fraud teams need to respond without placing unnecessary friction on the legitimate people they are trying to protect.