Utah girl tests ChatGPT, Claude & Gemini on 84 teen cases; finds empathy lacking

AI Summary

Eighth-grader Ellianna Gu evaluated ChatGPT, Claude, and Gemini using 84 scenarios involving common teenage concerns such as academic stress and loneliness. She found that the chatbots offered structured responses but often lacked empathy, and that response quality varied with prompt wording and format.

Ellianna Gu, an eighth-grader, tested AI chatbots including ChatGPT, Claude, and Gemini for emotional support. She designed 84 scenarios involving common teenage issues such as academic stress and loneliness. The chatbots provided organized responses but often lacked empathy, which is crucial for emotional challenges. The effectiveness of their replies varied based on the prompt's language style and format.