Automation Bias and Cognitive Biases in Financial Portfolio Allocation Decision-MakingPeng, Lauren FinanceAIAllocationUncertaintyFinanceB SThis thesis investigates how automation bias shapes cognitive biases and overconfidence in financial decision-making under uncertainty. As artificial intelligence (AI) becomes increasingly embedded in financial systems, from portfolio management to risk assessment, understanding how humans interact with algorithmic advice is essential. Through controlled experiments at Penn State’s Laboratory for Economics, Management, and Auctions (LEMA), participants will complete financial decision tasks under three conditions: independently, with a human advisor, or with AI advice. The study will measure decision quality, speed, confidence, and trust in AI to test whether automation bias mediates overconfidence and influences judgment quality. Building on behavioral finance, cognitive psychology, and human–AI interaction literature, this research aims to bridge theoretical gaps between algorithmic bias and human overreliance on technology. The results are expected to inform the design of ethical and transparent AI systems, as well as training programs that improve human–AI collaboration and decision accountability in financial contexts.Chloe Jeanne Tergiman, Thesis SupervisorBrian Spangler Davis, Thesis Honors Advisoropen_access2026-04-11T01:30:01Zhttps://honors.libraries.psu.edu/catalog/10323llp5314