Latin American criminal networks are consolidating power through the strategic application of artificial intelligence (AI), a shift profoundly transforming their activities and increasing their profitability. This evolution represents more than a simple adoption of new tools for illicit purposes; it signifies a structural metamorphosis. Operating within an increasingly automated, fragmented digital ecosystem where direct personal interaction is no longer required, criminal organizations have become more efficient, harder to track, and less exposed to traditional law enforcement risks.
“AI reduces costs, maximizes profits, and, above all, gives them anonymity. In many cases, the crime is no longer carried out directly by a person, but by an automated system. That makes it more difficult to identify and prosecute those responsible,” Juan Manuel Aguilar Antonio, a research professor at the Aragón Higher School of Studies at the National Autonomous University of Mexico (UNAM), told Diálogo.
The widespread dissemination of AI models such as GPT-4, Claude, and LLaMA, capable of understanding and generating text like a human being, answering questions and producing content in natural language, has granted advanced technological capabilities to groups that previously lacked specialized skills. Modern, user-friendly interfaces allow even technically unsophisticated criminals to launch high-level identity theft campaigns, execute social engineering (targeted psychological manipulation), and carry out highly personalized financial fraud.
Deepfakes and modern scam centers
A primary weapon in this new arsenal is the use of deepfakes — AI generated audio or video that mimics real identities with startling realism — to facilitate large-scale fraud.
In August 2025, the Rio de Janeiro police dismantled a scam against ride-sharing platform Uber involving nearly 2,000 fraudulent trips. Criminals used AI-generated images to create fictitious driver and passenger profiles. During these “ghost trips,” profiles were deleted before payment was processed, forcing the platform to reimburse fake drivers. Similarly, in Bolivia, also in 2025, the criminal group Clan San Roque — named after the prison of the same name in Chuquisaca where the operation originated — used AI to clone the voice of former Minister of Education Omar Veliz Ramos. By disseminating credible-sounding job offers on social media, the group generated illicit profits exceeding $720,000.
“AI allows criminal groups to do the same thing they did before, but faster, with fewer people and greater reach. For example, today they can send thousands of personalized fraudulent messages in minutes, clone someone’s voice to extort a family, or automate cyberattacks that previously required entire teams of specialists,” says Antonio.
At the same time, criminal groups use chatbots — automated programs capable of simulating human conversations in real time — and AI agents to track social media, messaging services, and websites to identify vulnerable or more easily manipulated individuals and attack them through targeted operations.
Crime-as-a-Service and the dark web
According to a 2025 report by Israeli cybersecurity platform Kela, the underground market for AI-assisted cybercrime grew by 200 percent in just 12 months. On the dark web — the hidden layer of the internet used for illicit trade — criminal platforms such as WormGPT, FraudGPT, Storm-2139, and the more recent Xanthorox AI have emerged, turning cybercrime into a fully structured, service-based digital market.
“These platforms operate under a ‘crime-as-a-service’ model,” Antonio explains. “They develop tools for fraud, ransomware, or the creation of false identities and offer them to third parties for a fee.”
According to the expert, major organizations like the Mexican cartels of Sinaloa and Jalisco New Generation (CJNG), are increasingly outsourcing their digital operations. While maintaining their traditional vertical structures, they rely on specialized tech brokers to implement algorithmic solutions, such as smart routing for trafficking, voice cloning for extorsion, and sophisticated phishing campaign to harvest financial data.
“Anyone with minimal resources can access sophisticated technology to commit crimes, which multiplies attacks, expands transnational reach, and makes it extremely difficult for authorities to identify and attribute responsibility,” Antonio explains.
A shared response: U.S. and regional cooperation
To counter this technological surge, a unified front has emerged between the United States and Latin American nations. Recognizing that AI-driven crime respects no borders, the United States has shifted its security assistance toward “digital justice” and high-tech capacity building.
The U.S. Department of State, through its Bureau of International Narcotics and Law Enforcement Affairs (INL), has significantly increased funding for cyber-forensics training. This support ensures that regional police forces can maintain a “digital chain of custody,” allowing AI-generated evidence to be successfully used in court.
A landmark in this cooperation was the 2025 signing of the U.N. Convention against Cybercrime. Supported heavily by the United States and regional leaders like Peru and Brazil, this treaty provides a legal framework for the real-time sharing of electronic evidence.
The training ground: ILEA San Salvador
Central to this shared response is the International Law Enforcement Academy (ILEA) in San Salvador. Established in 2005, ILEA has become the region’s premier “schoolhouse” for combating transnational organized crime. Through a partnership between the INL and the government of El Salvador, the academy hosts specialized courses addressing cyber threats.
ILEA has intensified its focus on Computer and Network Intrusions and Cryptocurrency Investigations. Instructors from the U.S. Secret Service and the FBI provide law enforcement officers from throughout Latin America with state-of-the-art tactics for “cyberexploitation” investigations. These programs are designed to evolve as quickly as the criminals do, teaching agents how to dismantle the digital infrastructure — such as the botnets and synthetic media servers — that underpins modern criminal enterprises.
Addressing vulnerabilities
Despite this progress, gaps remain. In April 2025, Peru, which already has regulations on AI, updated its Penal Code to include crimes involving the use of AI-based technologies. Argentina, Chile, Colombia, and Brazil, are still in the process of drafting and developing legislative proposals on the matter.
“The other problem is that institutions tend to be fragmented: The police, prosecutors, and financial agencies do not always work in a coordinated manner,” says Antonio. According to the expert, another factor that increases vulnerability is the wide technological gap between the state and criminal groups, which in many cases innovate and adapt more quickly than the authorities.
“In addition, we live in highly digitized societies, but not always with sufficient digital education. This makes it easier for people to become victims of fraud or manipulation,” adds Antonio.
Therefore, a comprehensive strategy is needed, not just a technological one. “Updated laws, greater cooperation between countries, and strengthened digital investigation capabilities are required. It is also key to invest in cyber intelligence and specialized training within the state,” Antonio adds, stressing the importance of protecting people’s rights. “Fighting crime with AI cannot mean restricting digital freedoms. The challenge is to find a balance: to develop technological sovereignty and response capabilities without sacrificing the rule of law.”
