Recent research from the University of Bath reveals that when lenders use Artificial Intelligence (AI) to make loan decisions, discrimination against women can worsen. However, the study also suggests that ethical lenders can tweak AI algorithms to reduce this bias while still boosting profits and enhancing their brand reputation.
The study focused on how AI influences loan decisions at car dealerships in Canada. It found that while AI significantly increased lenders’ profits, it did not worsen existing biases against marginalized groups—with the notable exception of women.
Increasing revenue
“We discovered that lenders could increase profits by applying machine learning to thousands of car loans. But this profit comes with a significant downside: AI can make car loan deals disproportionately less favorable for women, worsening social injustice in the process,” the researchers explained.
The research showed that using AI to optimize salesforce commissions could increase annual profits by up to 8%, but this came at a greater cost to women. However, the researchers found that by adjusting the AI algorithms, the bias against women could be reduced, still allowing for a profit increase of up to 4%.
By programming the AI to maximize profits while ensuring that women were not further disadvantaged, the study showed that lenders could still boost profits by focusing on non-marginalized groups—primarily men.
Smarter use
“This highlights the potential for AI to positively impact social welfare while also benefiting firms through immediate profit gains and long-term reputational advantages from socially responsible behavior,” the authors noted.
They point out that bias against women in this sector is well-documented. Previous research has shown that salespeople often negotiate less favorable loan terms for women than for men. Some studies suggest this could be because salespeople assume women are less knowledgeable about products like car loans or less assertive in negotiations.
“To address this injustice and avoid regulatory action, it might seem wise for firms to avoid AI altogether—but that would mean missing out on the profit boosts AI can offer. We argue there’s a middle ground: AI can be used responsibly to balance the trade-off between profits and social justice,” the researchers concluded.
“There’s no doubt that AI, if used thoughtlessly, can worsen discrimination against women. But if used responsibly, AI isn’t the threat it’s often made out to be. Rather than limiting AI, we should encourage firms to use it in a way that benefits both their business goals and social equity.”
