The survey, conducted by Censuswide, gathered responses from 200 claims handlers. Of the total respondents, 94% believed at least 5% of all claims had been fraudulently altered using AI. Furthermore, 19% suggested that AI-driven fraud was present in 11-25% of cases, while 64% suspected AI involvement in 5-10% of claims.

Manual claim reviews come with high costs, prompting some insurers to automatically approve lower-value claims. According to the survey, fraudsters appear to be exploiting this, with 93% of UK claims handlers suspecting smaller claims, particularly those valued between £501 and £1,000, to be most prone to the use of AI-generated or altered evidence, such as images, medical reports, and valuation certificates.

The increased accessibility of AI tools, coupled with the cost of living crisis, has likely contributed to this surge in fraudulent activity. Since the crisis began in 2021, 65% of handlers reported an increase in fraudulent claims, with 45% seeing a slight rise and 20% noticing a more significant escalation.

Commenting on the research and its findings, Roi Amir, CEO of Sprout.ai said: “Insurance fraud costs the industry billions, and this latest data shows it is on an upwards trajectory. Fraud has profound economic and societal consequences, driving up costs and premiums for insurance customers. The use of readily available AI tools to create and edit supporting documents for insurance claims is prolific, an issue that insurers need to regain control of by fighting AI with AI.  Powerful data processing tools, enabled by generative AI, can detect fraud by comparing the current claim against a vast database of imagery, documentation and existing claims data, flagging irregularities in real time.”