Today, it is possible to copy a person’s appearance using a simple photo and artificial intelligence, making facial recognition software a target for questioning. Financial institutions, marketplaces and healthcare companies, which rely on technology for security, are strengthening their systems with more layers of biometrics. The return of fingerprints is one of the protocols expected for next year.
Unico, which claims to serve 4 of the 5 largest banks and 9 of the 10 largest retailers in the country, recorded a 900% increase in episodes of sophisticated fraud (including the use of deepfakes) against identity verification, comparing 2025 data with last year.
“This trend reinforces the fact that it has become less costly for criminals to carry out this type of scam and reinforces the relevance of applying proof of life in facial recognition,” says Davi Reis, technology consultant at Unico. It’s a technology that analyzes everything from facial movements to blood flow under the skin to confirm the existence of a real person.
“With the evolution of threats, the trend is towards a new cycle of innovation, as is the case with the use of infrared to map the veins of the hands, which is booming in China,” explains the Unico consultant.
The competitor Identity.io arrived with a new proposal: save the analysis of the drawing formed by the taste buds, the furrows of the skin at the fingertips. Unlike the face, already exposed on social networks or on the WhatsApp profile, few people take high-resolution photos of their hands, explains the company’s executive director, Eduardo Ferrazzi.
According to him, Identity.io has already signed a contract with a major Brazilian bank and the technology is popular in Mexico and Central American countries. The company claims to be able to collect fingerprints with a photo.
To compensate for the lack of image quality, the company uses an artificial intelligence filter that complements patterns with your fingertips. The technical requirements for the camera and smartphone system requested by Identity.io are in line with what the banks’ own applications require in terms of security criteria, Ferrazzi explains. “We are not going to impose a technological barrier on low-income customers.”
Banco Popular Dominico, for example, began creating digital accounts with this technology in 2024. Registration time takes five minutes, according to data from the financial institution.
In Brazil, the lack of a fingerprint database for the entire Brazilian population poses a challenge. Serpro makes the images recorded when issuing driving licenses available, but those who do not have a license are exposed. Identity.io says it compensates for this gap with data stored on the device and maintained by the financial institution that will be its customer.
The requirements of the LGPD (General Data Protection Law) regarding the storage of fingerprints are the same as those required for facial recognition: the collection of biometric information requires special care.
Entities representing banks and payment institutions are increasing facial recognition security, combined with other layers of protection. To do this, they cross-reference mobile phone metadata (such as geolocation, connection network and device) and use artificial intelligence to block attacks on the code in real time.
In a simulation carried out for the LeafIdentity.io tested its own solution and showed that the deepfake image made by one of the software used in the report would be blocked. The blocking would be caused by the source of the image, a virtual camera created by the application, and not by the device connected to the cell phone or computer.
The app returns “unsecure camera feed” – the English expression for untrusted camera. The problem, according to Ferrazzi, is that not all financial institutions are using the latest techniques. “Today it’s not enough to ask the customer to move their face, it’s a previous generation technology.”
Unico, which still makes facial recognition one of its main methods, is developing its proof of life technology in Kazakhstan. “We block everything from more basic attacks, which use powerful tools developed for entertainment purposes, to advanced attacks, with techniques such as machine learning and expensive equipment,” says Reis.
Smartphone makers like Apple, Samsung, and Huawei are also beefing up their cameras with dedicated life-proof electromagnetic and infrared radiation sensors.
About The Author
