Credit scoring models that use AI may inherit biases from historical datasets which penalize groups who were previously denied credit scores more often and may overlook alternative factors that demonstrate credit worthiness.

• A case study report by the Student Borrower Protection Center found that a community college borrower would pay $1,314 more on a $10,000 loan than a student with the same credit profile pursuing a bachelors’ degree.

Predatory lending targets and imposes unfair or abusive loan terms on a borrower, often through deceptive or coercive actions. AI can identify targets and create content for these scams.

• 🏆In 2021, a report from The Markup showed how advertisers were targeting people for home equity loans, credit cards, and other financial services based on their age. Their work exposing this practice led to Facebook promising to take action against these advertisers.
• 🏆Google also began cracking down on predatory loan advertisements with fraudulent terms, preventing corporations from using personalized ad targeting for finance ads.

Insurance companies and banks can use models that take location, race, and other demographic information into account. These tools risk amplifying discriminatory outcomes that already harm vulnerable communities.

• A report by ProPublica and Consumer Reports explored how auto insurance rates for a Chicago resident in East Garfield Park was nearly 4 times as much as a Lakeview resident across town, despite lower vehicle crime rates in Lakeview. The analysis pointed to differences in rates for zip codes with larger Black and LatinX populations.

Bad actors can use artificial intelligence to create realistic fake identities to open accounts or conduct fraudulent transactions.

• In 2023, an investor in Florida was the target of an deepfake voice scam that mimicked his voice to trick his banker into handing over his money.