Year: 2026
Type: dissertation
Abstract: One of the greatest public health challenges of our time is mental health issues, particularly the increasing prevalence among young adults in Sweden. Meanwhile, artificial intelligence (AI), which is more accessible and widely used than ever before, is being discussed as a promising tool in mental health care. This report investigates the potential of a generative AI-based chatbot to promote mental well-being among individuals aged 18-29. The study was conducted as a clinical innovation project in collaboration with “Främjande av psykisk hälsa i Generation Z” which works together with young adults. Both qualitative and quantitative methods were employed, including an interview, a survey and a literature review. The findings contributed to a rule-based prototype illustrating a simplified version of a potential future chatbot. One of the prototype’s most significant functions is continuous detection of high-risk words in user input. The findings also identified essential requirements, including that AI should not replace human contact and that all provided information must be evidence-based. In addition, the chatbot should avoid “love-bombing” behaviours and minimize the risk of addictive use. In contrast to many existing AI services, the study suggests that co-design with young adults is necessary to increase engagement and willingness to use AI-based mental health support. Furthermore the results highlight the importance of security and transparency, as scepticism toward current services is common. Overall, the findings indicate that the identified aspects are essential for a future chatbot to be accepted and to fulfill it's intended purpose.
Authors: Gustafsson, Ellen, Rydberg, Lisa
Language: Swedish
Subfield: Applied Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7166263918