A User Study on Age-related Variances in the Perceived Emotional Alignment of Generative AIs

Objective: This study aims to assess whether different age groups of generative AI users perceive AI's emotional intelligence differently, especially in its ability to align with users' emotions.<br/>Background: Generative AIs have been rapidly implemented into users' lives, yet the potential and challenges of these models remain uncertain. We investigated how well these models understand users' emotions, specifically by age-related characteristics, which remain under exploration.<br/>Method: Participants from various age groups (20 to 60, N=283) participated in the study and evaluated generative AI artwork generated from specific prompt instructions to reflect the writer's emotions (Emotion-focused; EF). Then, these artworks were compared to the ones generated from prompt instructions that were to reflect only the facts about the story (Information-focused; IF). Then, participants rated 240 images based on how they align with the user's emotions reflected in the story.<br/>Results: Older adults (40s, 50s, and older) perceived AI more favorably in their alignment with user emotion than younger adults (20s and 30s). EF prompts outperformed IF prompts with significantly higher emotional alignment scores.<br/>Conclusion: Contrary to popular belief, older adults were more favorable to AIgenerated works than younger adults. Both younger and older adults preferred AIgenerated work that reflects human emotions over those do not.<br/>Application: Our findings highlight the importance of enhancing AI's emotional intelligence in increasing various demographic user engagement, especially in older adults. With increasing markets targeting senior users, these generative AIs can potentially increase accessibility and adoption rates in creative activities and businesses targeting well-being, healthcare, or education.