Booting Out Prejudice from Insurance

Accusations of unfair discrimination due to bad data entry have popped up in banking. We wanted to know what the insurance industry is doing to prevent that from happening to them. According to Tom Clay, who leads data science at Covéa, their teams spend a long time trying to establish what data is used when analysing a fraudulent claim. To avoid taking biases and prejudices into account when establishing a conclusion they're quick to look at the hard data they have, rather than external statistics. This provides a sensible baseline of data to work from and can avoid unnecessary slip ups. "When you start with people… that's where you get fair data and route out obvious biases and prejudices."