The Chair Cuatroochenta of Artificial Intelligence, Health and Welfare of the Universitat Jaume I has participated in the XXIII International Congress of Geriatrics and Gerontology of Vigo, held on June 9 and 10, to explain its projects and research on the use of technology for the detection of unwanted loneliness. Although it is not classified as a disease, it is a risk factor for the health of the elderly that can trigger other ailments, such as anxiety, depression or dementia, and may even be related to an increased risk of mortality.
13.4% of the Spanish population feels lonely, according to a study 'El coste de la soledad no deseada en España' (The cost of unwanted loneliness in Spain) carried out by the ONCE Foundation. Although the prevalence is higher in the younger population, unwanted loneliness in the elderly is linked to a greater use of the emergency service, primary care and a longer hospital stay. For all these reasons, addressing it is a priority for administrations and researchers. In fact, the Chair has been working for years on Serena, a conversational robot that can help detect situations of unwanted loneliness. The Chair's researcher, Mario Rojano, was in charge of presenting the project at a round table at the Congress in Vigo, entitled "Interventions to reduce social isolation and loneliness". His intervention focused on answering the question: Can technology reduce loneliness and social isolation?
Detect situations of loneliness with language processing.
Serena arises from a project funded by the Spanish Foundation for Science and Technology (FECYT). The aim of the project was to bring machine learning to citizens through a social problem such as unwanted loneliness in the elderly.
It is a chatbot that from the mobile phone, through a natural and daily conversation, questions the person with a questionnaire based on the Jong Gierveld Scale and the UCLA Scale for detecting loneliness. In addition, Serena incorporates open-ended questions that refer to feelings of loneliness, allowing us to collect information about different ways of feeling and expressing this feeling. The information collected allows us to train an AI-based model that learns the language patterns that characterize loneliness and non-loneliness texts to subsequently identify patterns where the response is in line with one or the other category.
Advantages of using Serena
Next steps with Serena
From the Cuatroochenta Chair of Artificial Intelligence Health and Wellbeing it is important to develop resources or tools that allow health professionals to detect if a person is at risk of loneliness so that they can intervene early or even prevent the adverse consequences related to unwanted states of loneliness.
«The use of new technologies can be of exponential help, allowing us to reach more people and increase the efficiency of detection and referral processes as well as help in interventions. Always bearing in mind that these technologies should act as a complement and not as a substitute for the relational and affective part of care and social-health care.»