Artificial Intelligence (AI) is emerging as a key technology in many sectors, including the fight against fraud. Thanks to its advanced analytical capabilities and its ability to detect anomalies, it offers innovative solutions for identifying and countering fraudulent activities. However, this same technology can be exploited for malicious purposes, facilitating ever more sophisticated fraud. This dual face of AI raises major questions: is it a reliable ally in the fight against fraud or a dangerous tool in the wrong hands?
The development of artificial intelligence is changing the face of fraud. On the one hand, it enables fraudsters to carry out increasingly sophisticated attacks that are difficult to detect (deepfakes, phishing). On the other hand, it enables better detection solutions to be created (analysis of transfer flows, risk scores, detection of unusual behaviour). Here’s an overview of the different challenges.
Artificial Intelligence as a fraud facilitator
The confusion between ‘compliance’ (‘control’) and ‘anti-fraud’ (‘detection’) methods is partly responsible for the rise in fraudsters’ skills. The fact that customers have to experiment with ‘controls’ allows fraudsters to test their methods. Because the fraudster will analyse, investigate and understand the different patterns in the customer journey. And therein lies the rub…
Systematically passing under a security scanner at an airport means understanding that you don’t have to wear metal to avoid a body search. It is therefore possible to adopt practices to create as little friction as possible and pass through the checkpoint quickly. This is what fraudsters do when they come up against checkpoints set up to combat fraud.
Adaptation by fraudsters… a relative adaptation :
Fraudsters only adapt to what they can see. They will always find a technology that will facilitate their fraud. Is the face-match a condition of control? They find out about the main principles and realise that face-swap gets through the checks most of the time. We also spoke to you a few months ago about the rise in power of Chat GPT and the generation of false documents, and AI is enabling them to become more proficient in their misdeeds.
Artificial Intelligence as a tool in the fight against fraud
To avoid a headlong rush to accumulate visible checkpoints and create an exhausting adoption experience for the majority of users, the solution is to position the majority of the fight against fraud in unobservable areas and/or based on a multitude of weak signals that cannot be deduced by iteration. The control points would be reinforcement points, to ensure, for example, that the user is indeed able to ‘confirm something’.
AI should be a means of countering fraud… but it should be used primarily as a tool in the fight against fraud. However, we must avoid using this formidable tool to beat fraudsters on their own turf. That is, on ground that they analyse, know and even master.
The key is to analyse the data using Artificial Intelligence
One of the challenges in the fight against fraud is to capture data and then extract information from it, which will lead to the development of knowledge about the phenomenon observed. Knowledge of the subject is the tipping point for determining the action to be taken. Applicable ‘wisdom’. This is the DIKW principle of Russel Ackoff’s pyramid, which dates back to 1989.
At Oneytrust, we believe that AI, as a support tool in each of these stages, is both the most effective and the wisest path.
English 
French 
Spanish 
Italian 
Dutch 
Portuguese 
Romanian