Designing transparent artificial intelligence systems to build consumer trust and accountability in financial decision engines

AI systems are increasingly being deployed to assist consumers in making complex and sensitive financial decisions. For example, financial services firms use decision support systems to assist with loan and mortgage payouts, wealth and asset management, investment and retirement planning, financial consultation, and credit risk assessment. The task of these decision support systems is often to estimate the model parameters and use them to suggest the best option. Since financial decisions are often very sensitive for consumers and mistakes can carry a huge cost for the consumer, these systems take the form of decision engines with decision support capabilities to flag likely bad decisions for consumers. Hence, the consequences of the output of these engines, in areas such as loan rejection or mortgage denial, can often have a large impact.