MIT Sloan professor Manish Raghavan is developing new approaches to measure and improve AI-driven decision-making systems, with significant implications for enterprise hiring practices and algorithmic fairness. His research examines how organisations can leverage AI's inherent observability to identify potential biases.
Raghavan's research challenges traditional assumptions about AI in hiring processes, noting that while AI systems inherit historical biases, they also present unique opportunities for improvement.
His work on medical decision-making systems provides valuable insights for enterprise AI deployment. Studying the Glasgow-Blatchford Score (GBS), an algorithmic screening tool, Raghavan found that while the system performs "roughly as good as humans on average," there are specific cases where human expertise remains crucial. This suggests a balanced approach to AI implementation, where algorithmic systems and human judgment work complementarily.
The research also examines how AI systems affect user engagement, which is particularly relevant for enterprises deploying digital platforms. Raghavan and his colleagues developed an award-winning model demonstrating how platform design can better align with long-term user satisfaction. "Long-term satisfaction is ultimately important, even if all you care about is a company's interests," says Raghavan, suggesting that companies can potentially improve both user experience and business outcomes simultaneously.
The research highlights opportunities for organisations to improve their algorithmic decision-making systems through better measurement and monitoring capabilities. Raghavan's work suggests that enterprises can benefit from AI's increased observability while maintaining crucial human oversight in critical decisions.
+