AI and Insurance fraud: new threats, new solutions

Artificial intelligence is transforming insurance fraud from the ground up. Fraudsters now have access to powerful new tools. But insurers do too, with unprecedented solutions to detect, qualify and eliminate fraud.

Key figures:

• 902 million euros in detected insurance fraud in 2024 (Alfa)
• 656 million euros in property and casualty fraud in 2024 (Alfa)
• 5 to 10% of claims paid may involve fraudulent cases, according to industry experts

What types of fraud will AI enable?

Generative AI gives fraudsters capabilities that simply did not exist before. Attacks are faster, more convincing and harder to catch manually.

Automated document fraud

AI tools can generate fake documents in seconds. Medical invoices, accident reports, claim supporting documents: everything can be falsified with near-perfect accuracy.

The most common document fraud types include:

• Fake bank identity statements and insurance certificates
• AI-generated accident reports with consistent GPS and weather data
• Synthetic medical prescriptions and reports
• Fictitious repair invoices and quotes generated on demand

Deepfakes and identity theft

Deepfakes allow fraudsters to impersonate a policyholder during a video interview or phone call. Identity fraud is now within reach of non-technical actors, with no particular IT skills required.

AI-Assisted organised networks

Fraud rings are already using AI to coordinate waves of consistent false claims. These "large-scale fraud" schemes target multiple brokers and insurers simultaneously, making manual detection virtually impossible.

What are the impacts for insurance stakeholders?

AI fraud does not hit every player in the same way, but it hits everyone.

For insurers, the risk is primarily financial and gradual. Undue claims weigh on combined ratios without always triggering immediate alerts: actuarial models process this data as ordinary claims. The problem is often identified with a delay, once the portfolio has already absorbed the impact.

For wholesale brokers, it is the quality of the underwritten risk that is at stake. Accepting a fraudulent risk upfront means carrying a liability that travels up the entire chain. Pressure on underwriting processes has never been higher.

For brokers, the challenge is more immediate: the client relationship. A proven fraud case in a portfolio means a damaged reputation and additional controls that add operational burden without generating value.

Overall, it is the honest policyholder who foots the bill through rising premiums, while the entire market faces growing distrust towards digital journeys.

How new portfolio management tools are using AI to fight fraud

Faced with these new threats, solutions like Korint are turning AI against fraud. These next-generation portfolio management tools operate on three levels.

1. Surfacing fraud

These platforms analyse in real time the data from claims, digital journeys and submitted documents. Anomalies are detected automatically: data inconsistencies, known document fraud patterns, suspicious behaviour during the underwriting process.

AI cross-references weak signals that human analysis cannot process at this speed and scale.

2. Qualifying and framing fraud

Once an anomaly is detected, some platforms produce a documented risk score alongside a clear explanation readable by business teams. Depending on the tool, this can help prioritise cases for investigation and adjust triggering thresholds by risk segment.

Capabilities vary significantly from one solution to another. Some wholesale brokers can configure their underwriting rules based on the signals flagged by the tool, while others use this data in a more manual way. The level of integration depends on the platform and the organisation using it.

3. Eliminating fraud over time

The deterrent effect is documented: AI detection platforms reduce repeat fraud attempts when combined with automated rejection actions. These tools allow action at the underwriting stage rather than at the point of claims settlement, which is the most effective lever for reducing structural fraud.

SourcesAlfa, fraud report 2024. Insurance Europe. France Assureurs, Annual Review 2023.