'Six Hours to Infiltrate a Bank': How AI Is Industrializing Fraud
With the constant development of artificial intelligence (AI), financial institutions in France and around the world are facing an unprecedented wave of risk. A little-known technological revolution that is set to intensify in the coming years, challenging our digital sovereignty.
AI as an Amplifier and Accelerator of Cyber Threats
One of the major issues affecting financial infrastructure today is that it is dependent on and "concentrated around a limited number of players." These players enable the functioning of the majority of payment and money transfer systems. If one of these third-party "players" is compromised by a cyberattack, the entire environment and all its clients suffer as a result.
Furthermore, access to vast amounts of information online or through data breaches allows malicious actors without significant technical or financial resources to carry out large-scale criminal operations. The financial sector, where stolen data is highly valuable due to its confidentiality, is particularly targeted. The AMF specifies that "64% of attacks carried out by cybercriminals [...] involved data breaches."
In February 2026, the French Directorate General of Public Finances indicated that the national bank account file, containing a list of banking details (RIB/IBAN), had been compromised by malicious actors.
$35 Million Stolen with a Fake Elon Musk
With this sensitive data, often purchased on the dark web, we are witnessing, according to the AMF, an "industrialization of fraud." Artificial intelligence absorbs your personal information to better deceive you and extract information or money. This is known as phishing.
The growing automation of AI does not only threaten financial institutions. In 2025, the AMF's investor assistance service received 11,262 scam reports, a 37% increase from the previous year. Today, individuals are also becoming prime targets.
With a savings rate that remains high in France this year (around 18%), young people are drawn to the emergence of banks and other "neo-brokers." These fully online applications offer savings solutions powered by massive promotional campaigns, primarily on social media. In 2025, "more than half of stock market transactions were made by investors under 35."
The recent broadening of the target audience logically facilitates fraudulent activities: 16% of French savers say they "have been victims of a financial investment scam," and surprisingly, young people are the most affected—32% among those under 35.
This phenomenon is manifested by the emergence of AI-generated deepfakes, "that is, the creation of video or audio content featuring a person for malicious purposes." In March 2025, Swedish television revealed one of the largest AI frauds. An international network based in Georgia stole more than $35 million from thousands of savers worldwide, using deepfakes of Elon Musk on social media to convince them to invest—never to repay them.
Having the Keys to Respond in Time
"We cannot prevent the trauma—that's a statement of lucidity," says Nicolas Arpagian, Director of Strategy at Jizô AI. In the majority of cases, criminals will have the advantage of mastering their target. They decide the modus operandi. When it is impossible to predict when an attack will occur, "resilience" must be the first response of victims. Acknowledging that an attack can happen, but having the keys to respond in time.
For the cybersecurity expert, "the future lies in the ability to observe in real time what is happening on the network. If you can tell who is connecting to what and for what purpose, a response scenario can be applied, if necessary." That is why artificial intelligence represents a solution for the financial sector and more generally for all sensitive economic sectors. It must be used to "detect what is normal behavior and what is abnormal behavior."
In the immediate future, national and European initiatives have been put in place to combat this wave of criminal acts, such as CERTs (Computer Emergency Response Teams), where financial sector players and institutions manage to communicate effectively to "document, characterize, and share information, even between commercial competitors, to better face future cybercriminal attacks."
Finally, training and risk prevention in all companies, large and small, remain effective measures. When someone is "entrusted with a computer or smartphone [...], they can be a gateway to malicious intrusion," says Nicolas Arpagian. In the future, everyone will likely have to face this threat risk, personally or professionally, and consequently know how to respond appropriately.
No comments:
Post a Comment