Uso de IA em fraudes de identidade dispara 300% e coloca e-commerce em alerta, revela Veriff

Digital fraud continues to put pressure on technology, finance and e-commerce companies on a global scale Identity Fraud Report 2026, which has just been published by Veriff, 4,18% of all verification attempts made in 2025 were fraudulent, equivalent to one in 25 digital interactions, keeping the level high for the third consecutive year.

The data gains relevance amid the growing sophistication of attacks. Compared to 2024, the use of digital media generated or manipulated by artificial intelligence grew 300%, indicating that AI has consolidated itself as one of the main catalysts of online fraud.

For Andrea Rozenberg, director of emerging markets at Veriff, fraud has become a structural problem of the digital economy.“The advancement of AI has reduced costs, increased scale and allowed highly sophisticated attacks to be executed much more efficiently”. 

As for solutions to prevent this type of crime, she warns that a single tool will never be enough to combat the multifaceted threat of fraud that exists in the digital world. “We are always aware of new behaviors and use artificial intelligence trained in global fraud standards that analyzes behavior, device and session context to identify risks.Sensitive cases also undergo human review, ensuring accuracy, security and regulatory compliance”, explains Andrea.

Latin America accelerates and lights up alert

In Latin America, data show a significant acceleration.In 2025, fraud attempts in the payments sector grew by 48%, while fintechs recorded an increase of 23.4%, reflecting the rapid pace of financial digitization.

Identity forgery, a practice in which fraudsters impersonate someone else to access or illicitly request digital accounts and services, accounted for 86% of attacks in the region. Passports, residence documents and national identities concentrate the highest rates of fraud, a direct reflection of the wide acceptance of these documents in digital processes. Although these tactics are known, artificial intelligence has significantly expanded the reach by enabling automation on an industrial scale and accelerating the creation of synthetic identities used to infiltrate digital systems.

In Brazil, data from the Veriff Fraud Index 2025, a survey based on consumer experience, reinforce the severity of the scenario. About 26% of Brazilians reported having faced fraud five times or more in the last 12 months, a percentage higher than that observed in the United States (15%) and the United Kingdom (10%). The financial impact is also significant: almost 40% of respondents reported having suffered losses above UST14T 251 (R$1 thousand).3 thousand losses in a single UST1.41 million).

E-commerce and finance lead risks

In 2025, e-commerce sites recorded a net fraud rate of 19.2%, about five times the global average, and high levels of authorized fraud.

The financial sector remained among the top targets, with net fraud exceeding 5.5%, a 30% index above the international average, driven mainly by AI-based attacks and deepfakes.

The scenario reinforces the need for continuous identity verification models, capable of combining automation, artificial intelligence and human analysis to mitigate risks in an increasingly digital and interconnected economic environment.

“We saw a trend in this direction in 2023 and 2024, but in 2025 the number of fraud attacks on e-commerce sites skyrocketed.Because largely they lack regulation and handle global transactions in the trillions of dollars, these sites are a tempting target for” fraudsters, warns Andrea.