How is APAC leading with AI to fight insurance fraud?

74% of respondents reported that fraud cases are either stable or increasing.

In Asia Pacific (APAC), the focus on fraud prevention is more advanced compared to other regions. About 78% of global respondents have fraud-specific teams, with APAC leading in the use of expert systems and machine learning to combat fraudulent claims. 

Fraud remains a major issue for insurers globally, with 74% of respondents reporting that fraud cases are either stable or increasing, RGA’s 2024 Global Claims Fraud Survey results showed.

About four in 10 of APAC respondents utilise these advanced tools, compared to 18% in the Americas and 12% in Europe, the Middle East, and Africa (EMEA).

A notable finding is the rise in AI-assisted fraud. The APAC region reported three times more cases of AI-driven falsified documents, such as death certificates, compared to the Americas. 

Despite this, AI also offers significant potential for detecting fraud, helping insurers combat these new threats.

Globally, the incidence of claims fraud has increased slightly to 3.8% from 3.6% in 2016. In APAC, fraud is a top priority, and the region has seen the most substantial growth in the deployment of specialised teams and technology to tackle the issue. 

Fraud investigations in APAC often centre on early claims and inconsistent explanations, with consumer and agent-assisted fraud being particularly prominent.

Looking forward, 68% of respondents worldwide expect fraud to increase over the next three to five years, driven by economic factors, digitization, and organised crime.

In APAC, insurers are well-positioned with advanced fraud detection tools, but the region remains a target for fraudsters leveraging technology to exploit the system. 

The survey was conducted online last March to April, with global claims professionals sharing their sentiments on fraud data from the calendar year 2023. The survey’s 83 respondents tripled the number of respondents from RGA’s 2017 fraud survey, which was based on data from 2016.