Artificial intelligence is starting to change everyday fraud in the Netherlands. Instead of obvious lies or low-quality forgeries, criminals can now generate realistic invoices, edit “damage” onto photos, write convincing complaint emails, and even imitate voices. Insurers and online retailers say this is making fraud faster, cheaper, and harder to spot.
Insurance fraud is becoming more sophisticated
Dutch insurers detected more than 9,000 cases of insurance fraud in 2024, according to figures from the Centre for Combating Insurance Crime (CBV), part of the Dutch Association of Insurers. The industry said it prevented €95.6 million in losses, which was almost €10 million more than the year before.
The concern is not only the number of cases, but the methods. Insurers say they are seeing more “smart” fraud that targets digital processes, including fake documents and manipulated images that look genuine at first glance.
How AI is used to fake claims
Recent reporting describes several common patterns:
Edited or generated photos that show damage that did not happen, or exaggerate real damage.
Fake receipts and invoices created in seconds, making it look like an expensive item was purchased.
False supporting documents (such as letters, statements, or other “proof”) that are formatted to look official.
Impersonation using AI-generated voice messages, where criminals pretend to be someone trusted.
These tricks are especially effective because many insurers and businesses handle claims and customer contacts online. When the process is digital, the first checks often rely on whether documents “look right,” and AI can now produce that look very well.
Retail and return fraud are also growing
The same tools are also helping fraud in e-commerce. Retailers can be hit by return scams, fake refund requests, and forged proof of purchase. While exact national totals for “retail fraud” are harder to measure in one clear statistic, Dutch reporting and industry messaging increasingly point to the same issue: fraud attempts are becoming easier to scale because AI reduces the effort needed to produce convincing “evidence.”
This matters to consumers too. When fraud rises, businesses often respond with stricter checks, slower refunds, and higher costs: changes that can affect honest customers as well.
Photo Credits: Sanket Mishra/Pexels
2025 data shows more scam reports and a holiday peak
Several 2025 indicators suggest the wider fraud environment is intensifying.
Fraudehelpdesk (first half of 2025): In a press release, the Fraud Help Desk said it received nearly 50% more reports in the first half of 2025 than in the same period of 2024. It also reported a strong increase in misleading sales practices and a very large jump in phone-based approaches.
The same document includes examples of rising categories. For instance, reports about fake webshops (valse webwinkels) rose from 2,120 (to 1 July 2024) to 3,177 (to 1 July 2025).
Aankoopfraude (purchase fraud) in 2025: SIDN, which manages the .nl domain, cited Fraud Help Desk data showing a 35% increase in purchase-fraud reports in 2025 (up to October), with reported losses of over €1.6 million—and warned that November and December often bring an additional spike because of holiday shopping.
AI voice cloning in scams (2025): Earlier in 2025, reporting based on AD and the Fraud Help Desk said nearly 10,000 scam call reports were filed in the first quarter of 2025 alone, describing schemes where criminals use AI to sound more credible.
Why the Netherlands is vulnerable to this shift
The Netherlands is highly digital: online banking is normal, online shopping is huge, and customer service is often handled through forms, chats, and uploads. That convenience is also what makes it attractive for fraudsters. When AI can mass-produce believable messages and documents, criminals can test many victims and systems quickly until something works.
The Dutch government’s Alert Online 2025 messaging also reflects this broader reality. It stated that almost three-quarters of Dutch people experienced attempts at cybercrime in the past 12 months, including phishing and increasingly “helpdesk fraud” by phone.
What insurers and businesses are doing
Insurers say they are also using AI and better data-sharing to detect fraud earlier, and they can take measures such as rejecting claims, ending policies, recovering investigation costs, or reporting cases to police.
At the same time, there is a clear balancing act. Stronger checks can stop fraud, but they can also slow down customer service and create friction for legitimate claims and returns. That is one reason many organisations are now focusing on “smart” controls: spotting unusual patterns, checking metadata, and verifying documents in ways that are harder to fake.
What this means for consumers
Most people will never commit fraud, but they can still feel the effects: more verification steps, slower reimbursements, and potentially higher premiums or prices if fraud losses grow. The bigger risk is that scams become more convincing, especially around busy seasons like Black Friday, Sinterklaas, and Christmas, when people are tired, rushed, and making more purchases.
