Reading:
AI Fraud Detection | AIlon in Risk
Image

AI Fraud Detection

Risk in Financial Services

  • Results aligned with science
  • Significant drivers of fraud identified
  • Fraud Hitrate 20x increased

The challenge

Insurance fraud is defined as exaggerating a legitimate claim. Science has shown us that the perception that insurance fraud is a crime committed mostly by “ordinary” and presumably otherwise respectable citizens is wrong.
We proved that AIlon can actually predict the likelihood of fraud based on psychometrics with astonishing results.

Results

Using AIlon, we identified a relatively clear pattern of psychometrics that lead to a high probability of insurance fraud. Indeed, AIlon found highly correlated habits to the probability of committing an insurance fraud, such as gambling, driver’s license suspension, and tax evasion. Other studies show a significant correlation of psychometrics, measured in the Big 5 taxonomy, and for example tax evasion: Tax evasion is significantly positive correlated to Openness towards new experience and significantly negative correlated to conscientiousness and agreeableness. One possible explanation is that people with high conscientiousness rationally perceive the possible risk, while individuals with high agreeableness tend to consider the decision morally wrong.
We have shown that self-control is indeed significantly correlated with the probability of committing insurance fraud, however in our population sample the best predictor by far has been as structural affinity for gambling.

Icon For Quote#1
Roughly 34% of the individuals who have a high probability of committing an insurance fraud are gambling, while just 4% of the people with a low probability do so.

Read the full story

Low levels of conscientiousness correlate with lower levels of sense of justice. This knowledge allowed the identification of high risk groups within the clients of a health insurance regarding fraud.
It is scientifically proven that self-control, as defined in the field of psychology, correlates with the probability of committing an insurance fraud. AIlon found that demographics actually correlate with demographics, such as sex or age, however AIlon also suggested that it is just a correlation and not a causality. Demographics correlate with crime, but they are not causes. AIlon's statements absolutely aligned with scientific crime research.
In the next step, ERASON applied the AIlon engine to investigate the correlation of Big 5, self-control and scientifically proven positive correlations such as gambling, license suspension and the probability of tax evasion.

Check the demo

No credit card required
Must be at least 8 characters