Colombia: AI, deepfakes and bank fraud

Incode warned that artificial intelligence is being used to create fake documents, generate synthetic images, and refine impersonation attempts against digital financial services in Colombia. The company also flagged risks tied to synthetic identities and deepfakes, while other outlets reported that these techniques are already being used to access financial services and open products with third-party data.

What fraud methods are being described?

Minuto60 described fraud methods in Colombia that include account takeover through social engineering and the opening of accounts or applications for credit using a real person's ID document together with a facial deepfake. Pulzo, for its part, reported that artificial intelligence is being used to create fake identities, forge documents and impersonate people for the same purpose.

In Incode's technical analysis, synthetic identities were defined as fraudulent identities built from a mix of real and invented data. The company also cited deepfakes as an attack tool and explained why network fraud detection matters in the face of agentic artificial intelligence.

What figures and scale indicators appear in the reports?

Pulzo reported, attributing it to Incode, that a single attacker may have been linked to 426 identity verification sessions, 35 altered documents and eight fintech companies in Latin America, although the source did not identify those fintechs or confirm that all of them were Colombian. In the same story, Pulzo attributed more than 765,000 complaints linked to fraud methods during 2025 to Asobancaria, with economic losses above 967 billion pesos.

The publication also said, citing Colombia's Superintendency of Industry and Commerce, that there have been 40,316 identity theft complaints since 2020. In another figure cited by Infobae, and also without verifiable methodology in the available result, 71% of Colombians would have faced at least one impersonation fraud attempt during 2025, in a scenario tied to vishing.

What are companies saying about the problem?

Incode said fraud detection at the network level matters in a context where artificial intelligence is speeding up document forgery, synthetic image generation and the adaptation of impersonation attacks. In its technical blog, published on Nov. 2, the company reiterated that synthetic identities blend real and invented data and that deepfakes remain an attack tool.

Portafolio reported, citing DataCrédito Experian, that 81% of credit originators use artificial intelligence. The same story said AI can enable new fraud scenarios, including deepfakes and impersonation, but can also be used for prevention and detection. That mix of uses has become one of the central issues now affecting banks, fintechs and verification providers in Colombia.