Colombia: Deepfakes in investment scams

Update November 3, 2026: the report adds new campaigns detected in Colombia involving fake ads that used deepfakes of David Vélez, Mario Hernández and Gustavo Petro to promote fictitious investments. It also adds recent data on cybercrime complaints, extortion with cloned voices and arrests linked to those cases.

Forbes Colombia documented 51 fraudulent ads published over 30 days, between March 10 and April 9, 2026, that used David Vélez deepfakes to promote a fake investment platform. The campaign included six unique deepfake videos and two active fraudulent domains spread across five Facebook pages.

What do these fake videos do?

These edits alter original recordings or generate new pieces with AI so they can pass as believable messages from known people. According to La Silla Vacía, a video attributed to a host and Colombian public figures was verified as fake, and the edited material directed users to pages where they were asked to provide personal data.

Warnings from Semana and El País Cali point to the same operating pattern, using a third party’s image to build trust and push the victim to fill out forms, invest in supposed products, or respond to a request for money. In both cases, the end goal is financial gain or the theft of banking data.

Forbes Colombia also reported a network with 24 fraudulent ads that used AI-generated videos of businessman Mario Hernández to promote a fictitious investment platform called Anticrisis. The campaign began on March 11, 2026 and, in addition to those 24 ads, Revelum reported they were distributed across 12 Facebook pages. A statement attributed to Hernández denied any connection to the content, platforms or offers.

Between January and May 2026, another 87 fraudulent ads circulated across 20 different pages and 17 different videos that used deepfakes of President Gustavo Petro to promote a fake investment scheme that appeared to have government backing.

What techniques appear in financial fraud?

A Semana report on financial fraud in Colombia argued that artificial intelligence can be used to impersonate identities, forge documents and defeat authentication mechanisms. The same coverage added that fraud detection can also rely on advanced analytics, with signals such as unusual changes in habits, unknown devices and new recipients.

The logic is preventive. Rather than waiting for the money to move, controls aim to detect anomalous patterns before the transaction is completed. The available material does not describe a single official Colombian response, but it does point to behavioral analytics as a key tool against impersonation attempts and digital fraud.

What other regional cases showed up in the sweep?

Tecnogus, citing researchers from ESET Latin America, described a regional investment scam network that used sponsored ads in Gmail, pages that mimicked news outlets and voice cloning with AI. That coverage recommended alerting the bank, preserving evidence and filing a complaint with cybercrime authorities.

The report does not confirm specifically Colombian victims or amounts, but it shows that the campaigns combine several layers of deception, from targeted advertising to voice clones. In parallel, DPL News reported, citing ColCERT, that criminals use AI to make campaigns more precise and harder to detect, and that fraud accounted for nearly 80% of the incidents handled by the team in the previous two years, although that figure could not be confirmed with the official document or the methodology used to calculate it.

As of December 2025, the National Police recorded 64 extortion complaints in Bolívar. Twenty-four were tied to digital methods that included cloned voices. GAULA recorded 36 arrests related to those cases.

Between January and March 2026, Colombia recorded 19,182 complaints for cybercrime. Theft through computer systems accounted for 10,328 cases, up 27.2% from a year earlier, and there were 891 complaints for website impersonation.

How are these campaigns moving in Colombia?

Recent cases show a more organized use of paid advertising and fake pages to push users toward fictitious investments. In the Forbes Colombia and Revelum records, the campaigns did not depend on a single video, but on dozens of ads distributed across multiple pages and domains.

That pattern appears in the impersonation of David Vélez as well as in the cases involving Mario Hernández and Gustavo Petro. In every case, the scam tries to build legitimacy with familiar faces, redirect users to a fraudulent platform and complete the capture before the victim spots the trick.

The available material does not confirm a single official Colombian response to all of these cases, but it does show that deepfake fraud is already combining extortion, voice cloning, targeted ads and impersonated websites in the country’s digital ecosystem.