"I'll just ask ChatGPT...": Conversational AI as a sense-making tool for physical and mental well-being
Authors
Ullerichs, Elia Hamann ; Fuglsang, Sofie Louise ; Jensen, Luna Annabell Linné
Term
4. Term
Publication year
2026
Submitted on
2026-06-18
Abstract
Conversational artificial intelligence (conversational AI) is increasingly used to address everyday concerns about physical and mental well-being. However, we still know relatively little about how people actually experience and understand these technologies in practice. This study investigates how people use and perceive conversational AI when dealing with well-being-related worries. We conducted semi-structured interviews with 20 participants who had previously used conversational AI for health-related questions or concerns. During the interviews, participants also completed a scenario-based task in which they interacted with a chatbot in a hypothetical situation. The interview data were analysed using reflexive thematic analysis, a method where the researcher systematically identifies key patterns and themes in participants’ accounts. Our findings suggest that participants often used chatbots as accessible, low-effort tools for dealing with uncertainty. Chatbots were used to explore symptoms, thoughts, and worries before these were clearly defined, helping participants organise information, consider possible explanations, and decide on next steps. Participants appreciated the speed, availability, privacy, and flexibility of chatbot interactions, but also described tensions around checking the information, feeling reassured, questions of responsibility, and trust in the system. The study provides an empirical account of conversational AI as a tool for making sense of everyday well-being concerns and shows how the very qualities that make these systems useful also shape how such concerns are understood and acted upon.
Samtalebaseret kunstig intelligens (konversationel AI) bruges i stigende grad til spørgsmål om fysisk og mental trivsel i hverdagen. Alligevel ved vi stadig ikke meget om, hvordan mennesker egentlig oplever og forstår disse teknologier, når de bruger dem i praksis. I dette studie undersøger vi, hvordan mennesker bruger og opfatter konversationel AI, når de skal håndtere bekymringer om deres egen trivsel. Vi gennemførte semistrukturerede interviews med 20 personer, der tidligere havde brugt konversationel AI til sundhedsrelaterede spørgsmål eller bekymringer. Som en del af interviewet gennemførte deltagerne også en opgave, hvor de skulle bruge en chatbot i en tænkt (scenario-baseret) situation. Interviewmaterialet blev analyseret ved hjælp af refleksiv tematisk analyse, en metode hvor forskeren systematisk identificerer centrale mønstre og temaer i deltagernes beskrivelser. Vores resultater tyder på, at deltagerne ofte brugte chatbots som lettilgængelige og enkle værktøjer til at håndtere usikkerhed. Chatbots blev brugt til at undersøge symptomer, tanker og bekymringer, før de var klart formuleret, og hjalp deltagerne med at strukturere information, overveje forskellige forklaringer og beslutte, hvad de skulle gøre som det næste skridt. Deltagerne satte pris på chatbottens hurtighed, tilgængelighed, mulighed for anonymitet og fleksible brug, men beskrev samtidig spændinger omkring behovet for at tjekke svarene, få reel tryghed, fordeling af ansvar og tillid til systemet. Studiet giver en empirisk beskrivelse af konversationel AI som et redskab til at skabe mening i hverdagspraksisser omkring trivsel og viser, hvordan de egenskaber, der gør disse systemer nyttige, også påvirker, hvordan bekymringer forstås og omsættes til handling.
[This abstract has been rewritten with the help of AI based on the project's original abstract]
Keywords