AI Profiling? Or Just a Stereotype? - How Machine Profiling Shapes Your Digital Identity

Explore how artificial intelligence systems develop emotional biases and create digital profiles that may stereotype users in this 33-minute conference talk. Discover the surprising finding that Large Language Models (LLMs) and AI agents exhibit emotional patterns that can override logical reasoning, leading to biased responses based on perceived user personas and minority characteristics. Learn about groundbreaking research from Harvard University and UC San Diego that reveals how AI systems develop hierarchical emotion organization, causing emotional patterns to dominate over causal reasoning in AI decision-making processes. Examine the implications of machine profiling on your digital identity and understand how AI's "secret labels" for users can influence the quality and nature of responses you receive. Delve into the research findings that demonstrate how stereotypes embedded in AI systems can significantly impact AI performance and user interactions, challenging assumptions about AI objectivity and neutrality.