Skip to main content

AI-CARES

AI information and chatbot assistance in well-being and mental health care

In this project, we explore the needs and possibilities regarding artificial intelligence (AI), such as ChatGPT, for organisations in well-being and mental health care. Together with those organisations, we want to safely take the accessibility, quality and efficiency of help and services to a higher level using AI. 

Date

From until

Supported by

VLAIO | TETRA
www.vlaio.be

View the project website

Project purpose

Digital help and services are increasingly an essential part of well-being and mental health care. People seek support and answers to questions via e-mail, chat, websites, FAQ sections, calls, social media and other digital tools. As in many other sectors, the use of artificial intelligence (AI) also offers great potential here. However, the accompanying challenges are additional. Indeed, qualitative, ethical and safe use of AI is essential when it comes to particularly sensitive aspects of our lives, such as well-being and mental health.

Therefore, we want to investigate and test how we can use AI to increase the accessibility, quality and efficiency of digital help and services in a safe and ethical way. By using AI as a support tool, we want to increase the reach of online offers in well-being and mental health care in Flanders. AI can help help seekers faster and more effectively and refer them to appropriate services. In addition, AI can help expand the online offer and help address existing pain points (such as long waiting times, language barriers and a shortage of volunteers).

Two tracks to achieve this objective

  1. Customised information delivery on well-being and mental health websites via an AI information assistant.
    A learning network is being set up, consisting of a large number of organisations from the broad sectors of wellbeing and mental health. Together, they are learning about the potential of conversational AI for tailored information delivery. After a needs analysis, a number of organisations within the learning network are effectively testing a virtual 'AI Assistant' on their website. Through AI-driven question clarification, tailored answers are formulated in the language of the help seeker. Answers are based on info and FAQ sections on their website, as well as external sources of information. Help requests that go beyond information clarification are actively referred to the online social map.
     

  2. Support for professional answering in chat and email help via an AI chatbot assistant.
    In anticipation of a chat conversation with a caregiver, an AI chatbot assistant can already request and process information so that the chat conversation can start immediately with the basic information gathered. Through AI-based answer suggestions, professionals or volunteers in online chat conversations can then respond to clients' needs faster and better. Chat conversations can also formulate answer suggestions in terms of question clarification and referral. Answerers then use these themselves in their conversations. In e-mail help, automatic suggestions of answers can be formulated, which are then finalised by the counsellor. Simultaneous translation into the language of the help seeker will also be made possible. In addition, we can explore how this AI assistant can be used in training, intervision and supervision of chat and e-mail helpers.

Team

Researcher

Sylvie Bernaerts

Sylvie Bernaerts (PhD) is a researcher in the People and Well-being research group, research line Psychology and technology. Her areas of expertise are immersive technology and artificial intelligence in (mental) health care.

Research line coordinator

Tom Van Daele

Tom Van Daele (PhD) is research coordinator Psychology and technology in the People and Well-being research group. As a clinical psychologist, he conducts research on the added value of technology within mental health care.

Related research lines