McLarens reports 300% rise in suspected fake document use in Q1 2023

Loss adjusting business McLarens UK and Ireland reported a 300% rise in the use of suspected fake documents in insurance claims in Q1 2023, with the firm’s head of investigation, Brendan Gillooly, attributing this uptick to the impact of generative artificial intelligence (AI).

Generative AI refers to artificial intelligence systems that can create content such as text, images or music by learning from existing data and generating new, similar outputs.

This type of technology includes deepfakes, for example, which is a video, audio or image that has been convincingly altered or manipulated to misrepresent someone as doing or saying something that was not actually done or said.

Gillooly told Insurance Times that there has been ”a notable jump in the number of false documents” used within insurance claims over the last 18 months, with generative AI making it increasingly difficult to detect these forgeries.

McLarens’ fraud investigation team has analysed a variety of suspected fake documents, including identity papers, machinery hire contracts, quotes, invoices, receipts, bank statements, deepfake images and audio recordings.

These alleged fake papers have appeared across a variety of claims, such as business interruption, commercial property, construction and engineering and private clients and estates.

The loss adjuster noted that it had also encountered several cyber losses involving deepfake video and audio recordings. For example, it handled one multimillion pound claim that involved payments which were fraudulently verified by a suspected AI deepfake.

Gillooly said: “Fakes are becoming more sophisticated and easier to come by.

”Today, there are dark web tools that let you purchase sophisticated fake documents for under $30 (£23). With the onset of generative AI, it’s becoming increasingly difficult to spot these false documents.”

Response

In response to this trend, McLarens has expanded its internal training programme.

The firm is also just starting a pilot of an AI document analysis tool, which has been designed to enhance its manual review and investigation process.

McLarens plans to collect data from the pilot over the course of 2024. At year-end, it will decide whether to permanently integrate the tool into its practices or extend the pilot period.

Gillooly, added: “Our response needs to be both technology and people led. Where we used to use electrostatic data analysis (ESDA) to allow us to see imprints on documents, graphology experts and ink test analysis to help demonstrate alterations in written form, we now rely on metadata analysis and specific software to detect digital manipulation.

“Our adjusting teams are the first line of defence and are crucial in conducting the necessary due diligence and spotting red flags.”