CSS Health Study Finds 42 Per Cent of People in Switzerland Used AI Chatbots to Check Symptoms

The share of people in Switzerland who use AI chatbots to make sense of symptoms has roughly doubled in a year, from 20 to 42 per cent, according to the 2026 CSS Health Study published on November 3. The research institute Sotomo surveyed 2,459 people in the German-, French- and Italian-speaking parts of the country for the health insurer. SRF and swissinfo.ch reported the findings under the headline that four in ten Swiss now use AI for self-diagnosis.

Use has risen in every age group. Among young adults aged 18 to 35, 62 per cent asked a chatbot about symptoms in the past year. Among people over 65, the figure rose from 10 per cent a year earlier to 30 per cent. More than a quarter of the population (28 per cent) now count AI chatbots among their most important sources of information on physical symptoms, up from 18 per cent, and for psychological symptoms the share rose from 11 to 16 per cent. Doctors, the internet and family and friends are still cited far more often, which CSS reads as a sign that chatbots serve as an extra source rather than a replacement for medical advice.

The study also points to a more personal role. Among people who feel lonely and use chatbots for private matters, 27 per cent said talking to a chatbot had sometimes eased their loneliness, and 18 per cent of those who use AI for private matters have told a chatbot something they have not yet told any person. At the same time, only 50 per cent of respondents turn to people around them when they are struggling psychologically, the lowest share since 2022.

What it means in practice

CSS Health Study Finds 42 Per Cent of People in Switzerland Used AI Chatbots to Check Symptoms — contextual photo

The jump puts more weight on how reliable these tools are, in health questions as in customer service, where a recent test found chatbots understand Swiss German but still invent customer promises. Insurers are also experimenting with their own systems, as in the on-premises AI pilot by Giotto.ai and Groupe Mutuel. The CSS study does not measure whether chatbot answers were accurate or what people did next.

What to watch is whether insurers, medical associations or the Federal Office of Public Health issue guidance on using chatbots for symptoms, and whether next year’s study measures how often chatbot advice leads people to see, or skip, a doctor.