Brazil: 42.5% of frauds already use AI

Brazil and the surge in deepfakes

In Brazil, 42.5% of financial fraud cases already use artificial intelligence tools, and deepfake use grew 830% between 2024 and 2025, according to a Finsiders Brasil article that attributes the figures to a presentation by a Federal Police agent.

The signal does not stand alone. LatamFintech reported that, according to Círculo de Crédito, the fraud rate in the fintech sector rose 64% in 2025 compared with 2024, moving from 2.2% to 3.6%. In that same regional picture, Yahoo Noticias reproduced a report saying that AI-driven identity fraud accounts for 48.3% of cases in Latin America, with Mexico City accounting for 13.85% of fraud alerts.

El Economista in Mexico published a broader version of that figure and added that AI-driven synthetic identity fraud reaches 48.3% of cases in Latin America. According to that report, banks and fintechs are seeing a significant rise in attempts to open accounts and request loans using fake data mixed with real information pulled from data breaches.

The impact on neobanks and fintechs

The El Economista report says attempts at synthetic identity fraud hit neobanks and consumer lending fintechs especially hard. The publication links that to fully digital onboarding flows and to reliance on document and biometric checks that can be bypassed with deepfakes.

In Colombia, ACIS described another defense line that is starting to gain ground against AI-driven financial fraud. The analysis says institutions are adding contextual data, such as geolocation patterns, device type, usage hours and historical transaction behavior, as a complement to biometrics.

ACIS also said traditional scoring and identity verification models in Colombia are being adjusted to detect anomalies in operating context, for example new devices or unusual locations. That shift comes in response to the rise of fraud carried out with synthetic identities and credentials stolen through automated phishing campaigns.