Artificial intelligence is not just transforming businesses, government administrations or creative industries. It is also giving cybercriminals the means to attack faster, at greater scale and with a level of sophistication once reserved for the best-organized groups. That is the finding of INTERPOL’s African Cyberthreat Assessment Report 2026, which estimates that AI is now involved in 55 percent of cybercrimes reported across the continent.
The organization describes a criminal ecosystem that has become truly industrialized and borderless, fueled by online scams, fake profiles, deepfakes and automated attacks. The financial losses recorded reflect this surge: they rose from $192 million in 2024 to $484 million, an increase of more than 150 percent. That figure likely represents only a fraction of the real damage, since many victims never file a complaint while businesses and institutions sometimes choose to stay silent to avoid a reputational crisis.
The report highlights above all a change in the nature of the threat. Complex technical skills are no longer required to run convincing scam campaigns: tools available to the general public can now draft a flawless fraudulent email, clone a voice, fabricate a video or create a nearly undetectable digital identity. Using generative AI, a scammer can produce hundreds of messages in seconds, tailored to the language, profession and profile of each victim, drawing on information gathered from social media: a professional post, a family photo, an announced trip or the name of a colleague is enough to build a personalized scenario. This shift particularly strengthens business email compromise attacks, in which criminals based in Africa manage to target companies in Europe or North America by impersonating an executive or a business partner.
The threat is no longer limited to the written word. Voice cloning can reproduce a person’s voice from just a few seconds of audio found online, turning any media-facing executive or active social media personality into a potential target. Deepfakes, for their part, are already being used to promote fake investments or fraudulent platforms, with the face of a well-known figure lending an appearance of legitimacy to the scam, particularly when the video circulates in a WhatsApp group or on a page mimicking a recognized media outlet.
Another finding in the report is particularly striking: the rise of synthetic identities. Criminals now combine authentic data, such as a name, phone number or official document, with entirely fabricated information, creating hybrid profiles capable of opening bank accounts, taking out digital loans or registering SIM cards while bypassing certain automated biometric checks. This vulnerability weighs heavily on the know-your-customer systems used by banks, fintechs and mobile money services, at a time when several African governments are accelerating the rollout of digital identity.
The INTERPOL document, based on responses from 36 African countries, found that 72 percent of surveyed states reported the presence of scam centers on their territory, with a particularly high concentration in Southern and West Africa. These operations sometimes function as genuine clandestine businesses, recruiting operators, dividing tasks and targeting multiple countries simultaneously through romance scams, fake investments, sextortion or account takeovers. The growth of mobile money creates especially fertile ground for these practices, with criminals exploiting the speed of transactions and the difficulty of recovering funds once transferred.
For INTERPOL, one of the major blind spots remains the lack of real-time information sharing between banks, telecom operators and law enforcement. One institution may detect a suspicious account without other relevant actors being immediately alerted, giving criminals time to disperse the funds. The organization also calls for better preparation among African investigators, judges and journalists in the face of synthetic content, deeming current readiness levels insufficient given how quickly criminal techniques are evolving.
Artificial intelligence, however, is not solely a weapon in the hands of fraudsters. It can also detect suspicious behavior, analyze large volumes of transactions and identify falsified documents. The battle is thus being fought between two uses of the same technology, and the advantage will go to institutions capable of combining technology, international cooperation and rapid response. The 55 percent threshold published by INTERPOL nonetheless marks a turning point: AI-driven cybercrime is no longer a future threat. It is already at work in the continent’s emails, phone calls, videos and financial systems.