The objective of SOCIAL HEALTH BOTS (2017-2021) has been to establish the knowledge needed to provide better and more efficient mental health services for young people (age 16-26 years) through smart use of chatbots, that is, digital machine conversation system which interacts with users via natural conversational language in an online environment.
The project has increased our understanding of the future potential of this type of technology, and the possible challenges of chatbot services in the context of mental health to support young people.
International studies reports that young people increasingly suffer from mental health issues such as social isolation, anxiety, eating disorders, sexual problems, depression, self-harm and suicidal thoughts.
This increase in mental health issues, and suicidal behavior combined with limited number of qualified health care staff, is a major challenge to the health care system and to the society at large.
The latest developments in artificial intelligence and in-depth learning, combined with the fact that many people use the internet for health information, as well as the emergence of large data sets, make chatbots a promising supplement to more labor-intensive interventions within health and welfare services.
The project has developed several different prototypes of such chatbots aimed at young people seeking health-related help. In addition, we have designed a chatbot prototype to support the work of health nurses in the school health service.
Many of the prototypes in the project have been evaluated, both related to technical response precision and user experience related to various forms of social support. When it comes to response precision based on machine learning, there is still a long way to go to 100% precision, partly because better training data is needed. However, several of the prototypes have received good feedback in the user evaluations.
When it comes to user experience several young people (aged 16-20) have tested out chatbots in the project over weeks and some even for months. These participants reported chatbots as a useful source to social support Social chatbots are also perceived as an efficient service to access trusted information support. The chatbot is perceived as a low-threshold service for mental health help for young people, especially because it is available regardless of time and space.
In more detail, young people, in our studies, report on chatbots as a source of: 1) emotional support (e.g. receiving care and empathy), 2) appraisal support (e.g. feedback and assessment on self), 3) instrumental support (eg. practical help), and 4) informational support (e.g. good advice and tips). This shows that chatbots can have an enormous potential not only as a source of social support in everyday life, and thereby have a preventive mental health effect.
Chatbots may help young people with mental health problems. Many young people report greater openness when they interact with chatbots, primarily because they feel that chatbots do not judge them. Young people see chatbots as a more anonymous way to get health-related information than through Google or phone-based help services. Since the chatbot is a machine, it is perceived to be good at keeping secrets and not sharing information with others.
In another study, we found that chatbots can be a useful supplement to the health service in general and especially within the school's health service. This study provides the necessary knowledge for chatbot developers and healthcare providers regarding (a) the support young people need from a chatbot in school healthcare and (b) how chatbot conversations can be designed to provide such support.
The proposed framework describes the different types of support a chatbot can provide, and will provide a useful background for chatbot developers and healthcare providers when working towards improved chatbot solutions.
However, there are limitations that need to be addressed before we can adapt chatbots in the mental health services targeting young people. Chatbot technology still has a long way to go before it can understand the user properly. It is also challenging to develop chatbots with privacy-proof solutions. In addition, it is necessary to deal with difficult situations such as suicide risk and reports of physical and / or psychological abuse.
We need more research to solve these problems if chatbots are to take on the role of a communication partner in the health sector and interact with young people on sensitive topics and mental problems.
The project's research partners have been SINTEF, with the research group Human-Computer Interaction, the University of Oslo, and the University of Agder, with the Center for AI Research.
Vi har løftet forskningsfronten på chatbots i Norge og internasjonalt. En rekke publikasjoner er hyppig sitert. Vi har også mottatt HONORABLE MENTION AWARD på den prestisjefylte konferansen CHI`21 ACM Press (https://chi2021.acm.org).
Vi har utviklet et nettverk at internasjonale chatbot-forskere gjennom den årlige og internasjonale workshopen Conversations.
Flere aktører i Norge har henvendt seg til oss om erfaringer og kunnskap knyttet til chatbots.
Vi har samarbeidet med BUFDIR for å løfte sin satsning på bruk av chatbots for å nå unge mennesker gjennom et nytt grensesnitt.
Vi har gjennom en rekke kronikker, innlegg og seminarer satt algoritmer og kunstig intelligens på dagsorden.
Flere protoyper av chatbots er blitt utviklet og testet. Ingen av disse er foreløpig blitt implementert, men har gitt nødvendig erfaring for å etablere viktig kunnskap om hvordan teknologien virker i møte med mennesker og samfunnet.
Findings from relevant international and national studies suggests that young people increasingly suffer from mental health issues such as social isolation, anxiety, eating disorders, sexual problems, depression, self-harm and suicidal thoughts. SOCIAL HEALTH BOTS address the increasing need for services that can help young people (16-26 years) with mental health issues.
Internet-delivered health services may be particularly suitable to reach the present digital generation of young people. The recent developments in artificial intelligence and deep learning, coupled with the broad popular uptake of online digital devices and the emergence of huge data sets (big data) make chatbots a promising supplement to more labour-intensive interventions within health and welfare services. However, the automating power of chatbots in the context of health services are largely understudied. SOCIAL HEALTH BOTS will, therefore, increase our understanding of the future potential of this type of technology, and the possible challenges of robotic services in the context of mental health.
Moreover, SOCIAL HEALTH BOTS addresses the need for research and innovation to improve and integrate public healthcare services. This by and understanding of how ground-breaking chatbots can supplement, coordinate and improve relevant health care services targeting young people - to facilitate information and social support for early intervention and a long-term healthy behaviour.
A human-centred design approach will support a deep understanding of relevant target users (e.g. young people, information providers, health care workers at schools, psychologists and doctors). Strong user involvement throughout the chain of research and innovation, including The Norwegian Directorate for Children, Youth and Family Affairs (www.ung.no), the health care services at schools in Oslo (Skolehelsetjenesten), and Save the Children, will enhance the relevance and uptake of the project result.