AI Chatbots Often Lack Empathy in Emotional Scenarios
A 2026 student project shows models offer structured, clear advice while falling short on the empathetic connection teens seek.
Updated on Oct. 24, 2026 in Artificial Intelligence
Live Poll
Do you trust AI chatbots to provide helpful emotional support for your personal problems?
In 2026, Salt Lake City eighth-grader Ellianna Gu analyzed how three AI chatbots respond to 84 emotionally challenging scenarios. Her project, which evaluated performance based on empathy, clarity, problem-solving, and helpfulness, found that chatbot responses often lacked empathy.
Why it matters
As more teenagers utilize AI platforms for companionship and emotional support, understanding the limitations of current large language models in handling sensitive social or personal scenarios is critical for evaluating their role in adolescent mental wellness.
The study utilized a standardized set of 84 scenarios across seven emotional categories, including loneliness, anger, and social anxiety. Evaluation was performed against four specific criteria: empathy, clarity, problem-solving, and helpfulness, with Claude scoring a three for effectiveness.
The players
Ellianna Gu
An eighth-grade student based in Salt Lake City and a 2026 finalist in the Thermo Fisher Scientific Junior Innovators Challenge.
OpenAI
The developer of ChatGPT, a large language model based on the transformer architecture widely used for generative text tasks.
Anthropic
The creator of Claude, an AI system designed with a focus on constitutional AI and safety standards in conversational interactions.
The parent company behind Gemini, a multimodal AI model trained to process and integrate text, images, and video information.
The details
Gu created realistic emotional situations using AI to generate the scenarios before manually and algorithmically assessing how the models responded. She observed that performance on these benchmarks varied significantly depending on whether the input prompt utilized formal language or slang. The assessment specifically measured the models' ability to provide structured and actionable advice against their inability to demonstrate emotional resonance.
Timeline
2026: Ellianna Gu became a finalist in the Thermo Fisher Scientific Junior Innovators Challenge.
The Tech Race
This project contributes to the ongoing evaluation of how large language models function as digital counselors. It adds to a growing body of research benchmarking AI utility against human social needs, a space where companies like OpenAI, Anthropic, and Google are competing to improve reasoning and tone.
Users seeking emotional support should remain aware that current AI models may provide structured responses while lacking consistent empathy. Performance can fluctuate based on phrasing, suggesting that teenagers and other users should use caution when relying on AI for personal advice.
The takeaway
This study highlights that while AI is increasingly capable of structuring complex advice, it currently struggles to meet human standards for empathy. Interested readers should watch for future longitudinal studies measuring the effect of AI interactions on teenage social development.
Further reading
Explore deeper analysis on the current state of Artificial Intelligence and its social implications.
Source note: This article includes information reported by The Times of India.
Live Poll
Do you trust AI chatbots to provide helpful emotional support for your personal problems?
