Major AI Models Systematically Recommend Pricier Options to Wealthier Users, Study Finds

New research reveals that leading large language models from OpenAI, Anthropic, Google, and Qwen exhibit systematic bias in product recommendations based on inferred user wealth, steering shoppers toward more expensive options even when explicitly instructed otherwise. Researchers from Cisco Foundation AI and Carnegie Mellon University conducted 325,000 experiments across 13 AI models, testing their recommendation behavior across three high-stakes economic domains: airline flights, health insurance, and graduate school programs. The results demonstrate a troubling pattern: eight of the thirteen models consistently recommended pricier options to wealthier user profiles when presented with identical purchase requests.