Labeling in Their Shoes - Improving Text Annotation with Cognitive Empathy Priming

Explore how cognitive empathy priming can enhance the quality and consistency of text annotation tasks in this 18-minute conference presentation from the Association for Computing Machinery. Discover research findings on how encouraging annotators to consider different perspectives improves labeling accuracy and reduces bias in machine learning datasets. Learn about the methodology used to implement empathy priming techniques and examine the measurable improvements in annotation quality when human labelers are prompted to think from multiple viewpoints. Understand the implications for creating more inclusive AI systems through better training data, as researchers Sung Hyun Kwon, Jessica Clark, Il-Horn Hann, and Jui Ramaprasad present their work on bridging cognitive psychology and machine learning annotation practices.