Deepfakes threaten the economy through financial fraud. Companies must strengthen detection, training and adopt increased vigilance to limit risks, using AI.
The images and videos generated by AI represent a growing threat to society and the economy. They are more and more easy to produce and much more difficult to distinguish from reality. Until now, the debate has mainly focused on political -aimed Deepfakes, used by malicious actors to try to influence democracies. However, in Europe, in the United Kingdom, in the EU and in India, these fears have proven to be largely unfounded. However, the data indicates an explosion of deepfakes worldwide, with an increase of 300 % in 2024. In France, this figure reached almost 140 % in comparison with the previous year.
From now on, concern focuses on a much more concrete danger: the use of Deepfakes in financial fraud, which threatens both businesses and individuals. If companies are aware of the risks linked to the identification of new customers – where the identity verification and the monitoring of fraud are essential stages – they remain largely helpless in the face of the scams by phishing and identity theft doped at AI. In the United States, identity theft was the first category of fraud in 2023, resulting in $ 2.7 billion in declared losses, according to the Federal Trade Commission. And the more these techniques perfection, the more the victims multiply. An awareness nevertheless takes place among managers: according to a Deloitte study, 15.1 % of the executives questioned have already been faced with at least one incident of financial fraud involving the use of Deepfakes in 2023, and 10.8 % with several.
Deepfakes and fraud: an increasing risk, a response still too low
The phenomenon only intensifies: more than half of the managers questioned (51.6 %) expect an increase in the number and extent of the attacks involving Deepfakes aimed at financial and accounting data. However, few concrete measures are taken. Worse still, a fifth of the respondents (20.1 %) admit that they have no confidence in their ability to react effectively to these fraud.
If Deepfake detection tools play a key role in preventing external fraudsters from bypassing integration verification processes, companies must also protect themselves against internal threats. An approach to “low confidence” in financial requests or critical decisions, combined with new digital tools boosted by AI, becomes essential for detecting swords by phishing and identity theft. This implies much more than a simple technological adoption: training, education and an overhaul of our approach to visual and audio content is necessary. This change must be driven at all levels of the organization, from the summit to the operational teams.
A holistic strategy of Deepfake
Sociocultural improbabilities: the context and logic may be the best tool for fighting “Deepfake” fraud. Each stakeholder, at each stage, must consider information with new skepticism. In the recent case where an employee of the financial service paid $ 25 million following a video call with a fake financial director, one wonders why the financial director requests 25 million dollars and to what extent this request comes out of the ordinary. This is certainly easier in some contexts than in others, because the most effective fraudster will design its approach so that it seems to correspond to the normal behavior of a person.
Companies must adopt generalized skepticism, subjecting videos and calls to the same checks as emails. Training employees and encouraging a double validation of audio and visual content makes it possible to better identify inconsistencies. Observe the biological signs – as well as the blinking of the eyes or the natural movements of the throat – rests an effective means of detecting anomalies. Breaking the usual patterns by asking for unusual gestures or by modifying the lighting can also expose manipulations. Finally, tools like for Fakes’ Sake for the image or Pindrop for the audio help to detect Deepfakes, although no method is infallible in the face of their rapid evolution.
Systematic skepticism in the face of Deepfakes
In the era of mass synthetic information, companies must apply the same level of vigilance to visual and sound content as to new contracts, the integration of customers and the filtering of illicit actors. Faced with internal and external threats, the adoption of AI -based verification tools, coupled with reinforced training and awareness programs, is essential to limit the financial risks linked to Deepfakes.
