How AI Is Now Also Confusing Insurers
Amsterdam, zaterdag, 22 november 2025.
In 2024, over 9,000 cases of insurance fraud were identified — more than ever before, largely due to the rise of artificial intelligence. The most striking development? Criminals are using AI to create convincing fake invoices and manipulated damage photos, sometimes even with voice messages that sound like genuine executives. What did the Dutch Association of Insurers reveal? The same technology enabling fraud is also being used to detect it. The battle is not only technological, but also a race in speed and cooperation. And the impact on your own premium? It’s greater than you think.
The Rise of AI in Insurance Fraud: From Fake Invoices to Deepfake Voice Messages
In 2024, over 9,000 cases of insurance fraud were confirmed, a rise of more than a thousand incidents compared to 2023 [1][2][3]. The emergence of artificial intelligence (AI) is cited as a key factor behind this increase [1][4][5]. Criminals are using AI to generate highly convincing fake invoices, such as one for an expensive Cartier sunglasses that was AI-generated [2]. Additionally, damage photos of vehicles are being manipulated or fully AI-created, making insurance claims appear more credible [2]. There have also been cases where AI-generated voice messages were used to mislead insurance staff, with messages convincingly mimicking those of senior executives [2]. Insurers suspect these fraud techniques are spreading due to the accessibility of advanced AI tools, which allow professional-looking documents and audio to be produced quickly [5]. The Dutch Association of Insurers has not provided detailed information on specific techniques or the number of cases per type, as they do not wish to ‘give potential criminals more insight’ [2].
Countermeasures: How Insurers Use AI to Detect Fraud
While criminals use AI to commit fraud, insurers are increasingly leveraging AI themselves to detect such fraud [1][2][5]. Advanced algorithms are being deployed to identify unusual patterns, such as a sudden spike in claim submissions within a short period, or atypical behaviour in digital application processes [2]. Insurers also use AI to detect deepfakes and manipulations in images or audio [2]. A key strategy involves requesting multiple photos of damage from different angles — a task still difficult for AI to consistently and convincingly replicate [2]. Furthermore, insurers are adopting a more integrated approach: they are sharing more information about dangerous patterns and detected fraudsters across companies, reducing the chances of successful fraud across multiple insurers [1][4][5]. This collaboration is facilitated through the External Referral Register (EVR), which recorded 3,567 fraud incidents in 2024 [1][3][4][5].
The Risk of Cyber-Related Fraud and the Role of Cooperation
According to general director Richard Weurding of the Dutch Association of Insurers, many modern fraud threats are cyber-related and becoming increasingly sophisticated [1][4][5]. The combination of AI and cybercrime poses a growing risk, with fraudsters attempting to steal internal insurer data through deceptive communication, such as emails or phone calls that appear to come from senior management [2]. To combat these threats, cooperation is crucial. In 2024, the Centre for Combating Insurance Crime (CBV) shared 258 cyber incident advisories with the sector, processed 3,138 incident reports, and handled 2,033 inquiries from insurers and law enforcement [1][5]. The CBV also received 96 anonymous tips from members of the public who suspected fraud [1][5]. This knowledge exchange helps insurers respond swiftly to new fraud tactics and prevents criminals from applying the same schemes across multiple insurers [1][4][5].
Impact on Premiums and Trust in the System
The rise in fraud has direct consequences for the premiums consumers pay. Due to the identification of 9,070 fraud cases in 2024, insurers saved €95.6 million, nearly €10 million more than in 2023 [1][2][3][5]. While this amount represents a substantial financial saving for the sector, it also underscores the scale of the damage caused by unauthorized claims [1][5]. According to the Dutch Association of Insurers, it is vital to fight fraud to prevent ‘honest customers from unnecessarily paying higher premiums’ [1][4][5]. When fraud is confirmed, insurers take strict measures, including rejecting claims, terminating policies, recovering investigation costs, and reporting to the police [1][2][3][5]. In 2024, 3,567 fraudsters were registered in the joint alert system, the EVR, significantly limiting their future access to insurance [1][3][5].
The New Challenge: Human ‘Gut Feeling’ as the First Line of Defence
Despite advances in technology, the Dutch Association of Insurers emphasizes that the human ‘gut feeling’ cannot be replaced by technical tools [2]. Staff are encouraged to remain vigilant: they must be alert to unusual signals, such as atypical communication or uncomfortable details in a claim [2]. The association warns that anyone within an insurer could face ‘sophisticated attempts by malicious actors through individual email, app, or phone traffic’ at any time [2]. The new policy vision ‘Resilient and Alert’, launched in November 2024, highlights the need for innovation and awareness across the sector, particularly in light of the rise of AI [5]. Insurers continue to invest in both technology and human capacity to sustainably combat fraud [1][2][5].