Fleet software provider FleetCheck has warned that rapid advances in artificial intelligence (AI) have increased the risk of fraudulent insurance claims for fleets, as image generation technology becomes more sophisticated.
Neil Avent, chief technical officer at FleetCheck, said the emergence of generative adversarial networks (GANs) brought a significant improvement in the realism of AI-generated images and video.
Avent said: “There has been a real step change in the last few months.
“GANs technology, which pits two neural networks against each other to generate convincing images, is now widely available via mainstream image-generation platforms.
“It has replaced previous diffusion-style systems and produces much more realistic results.”
This leap in image quality could make it easier for individuals to exaggerate the extent of vehicle damage following an accident.
Avent added: “If an employee is involved in a collision, there is a risk that the other party could use AI to manipulate the images and exaggerate the damage – for example, by adding scratches or dents to door panels.
“One of the strengths of GANs is that it can apply such alterations consistently across multiple images.”
While such fraud may not yet be widespread, Avent warned that AI tools are developing quickly and could reach a tipping point in the near future.
He said: “We’re approaching a stage where even experts may struggle to detect that an image has been altered.
“Fleets should be alert to this growing risk, particularly given the potential impact on insurance costs.”
To help counter the threat, FleetCheck encouraged fleets to adopt simple procedures that can support genuine insurance claims.
Avent said: “Drivers involved in an accident should be asked to take photographs of all vehicles from multiple angles.
“Our app supports this by prompting drivers through the process, helping to create a complete and verifiable record of the incident.”
Nick Hutson, compliance and training manager at 1st Choice Insurance, added: “Fraudulent and exaggerated claims have been an issue for decades, but the sophistication of methods is increasing.
“It’s highly likely we’ve already seen AI-manipulated images without knowing it.
“Encouraging drivers to take thorough, clear photos and capture accurate information at the scene can play a major role in identifying and preventing fraudulent activity.”
