Administrative Burden and Algorithmic Equity: A Randomized Trial of AI versus Human Assistance in Public Benefits Enrollment

More than $100 billion in allocated public benefits go unclaimed annually, due, in large part, to administrative burdens. Artificial intelligence (AI) may be able to reduce these burdens at scale. However, we lack rigorous evidence on AI’s ability to do so. Sociologist Steven Lauterwasser investigate how AI assistance compares to human assistance on reducing administrative burdens when enrolling in public benefits. He will conduct a randomized controlled trial for his study.