This originally ran in MarketWatch
Company expands its "Fairness-as-a-Service" solution to reduce algorithmic bias in the
LOS ANGELES, Aug. 23, 2022 FairPlay, the world’s first “Fairness-as-a-Service” solution for algorithmic decision-making, today announced the launch of Input Intelligence, a bias detection product for insurance carriers. This is an expansion of FairPlay’s existing fairness solutions for the financial services industry.
Algorithms are increasingly being used to make high-stakes decisions about people’s lives in insurance, lending, employment, and other areas. FairPlay’s new Input Intelligence tool will help insurance companies review their data for bias, ensuring they are not proxies for protected characteristics such as race or gender. Insurance companies will benefit from reaching more customers, especially from historically disadvantaged groups, and meeting growing consumer and regulatory demands for greater fairness in insurance.
FairPlay’s insurance division will be led by Josh Hershman, former Deputy Commissioner of the State of Connecticut Insurance Department. During his time as Deputy Commissioner, Hershman worked with stakeholders to establish guidelines around the use of big data, AI/ML models for insurance providers in the state of Connecticut.
“Consumers are demanding a better experience from insurance carriers, and the industry is meeting those demands using AI/ML at every stage of the consumer journey,” said Hershman. “Data variables used correlate to risk but may cause disparities for historically disadvantaged groups. AI/ML in insurance must be executed fairly and ethically, not only to meet consumer demand but also to give insurers better reach into underserved communities. I am thrilled to team up with FairPlay and work with carriers and regulators to alleviate this problem.”
Input Intelligence will also draw on FairPlay’s redlining technology, which identifies unfair practices based on where people live so that insurance carriers can ensure their marketing and products are not excluding low- and moderate-income neighborhoods.