Search

Purpose In the European-Taiwanese project Horizon AAL (HAAL), a dashboard is being developed that can act as a decision-aid for caregivers of community-dwelling people with dementia (PwD). In the dashboard, which is driven by artificial intelligence (AI), the data from a number of interoperable AAL (Active and Assisted Living) solutions are gathered and analysed in order to provide insights and predictions about the health and well-being of the PwD. In addition, the dashboard may provide recommendations that help caregivers to assess the care and support needs of their clients. The increasing advancements of AI-technologies such as the HAAL AI-driven dashboard come with benefits such as faster, more accurate, and more efficient data-analysis and the augmentation of human decision-making (Hassani et al., 2020), but also with challenges from a social and ethical perspective, e.g. related to privacy, transparency, human control and trust. In this line, it is broadly acknowledged that the proper embedding of healthcare technologies driven by AI requires innovators and other stakeholders to actively anticipate and reflect on, and be responsive to promises and risks and to societal values, needs and expectations (Morley et al., 2019; Tsamados et al., 2021 ; WHO, 2021). In the ongoing HAAL study, it is therefore explored what decisions and actions can be taken in the design of the dashboard and its cross-cultural implementation in order to account for the needs and values of end-users, and to achieve responsible innovation (RI) that is socially desirable, ethically acceptable and sustainable (Von Schomberg, 2013). Method Through a mixed-methods approach, a survey, focus groups and semi-structured interviews were performed in the Netherlands, Italy and Taiwan, to explore the perspectives of (1) HAAL project partners, (2) end-users and (3) experts outside the consortium on RI in HAAL. Two scenarios about the HAAL dashboard, with differing degrees of autonomous decision-making and learning by AI, were used to help respondents reflect on the potential positive and negative impacts and the values at stake when deploying such an AI-based application. Also, we explored the respondents' views about the decisions and actions that could be taken in the design and deployment of the dashboard to foster RI. Finally we compared the views of Dutch, Italian and Taiwanese respondents, thereby exploring how to be sensitive to both the local embedding and the wider applicability of the technology. Results and discussion As argued by ÓhÉigeartaigh et al. (2020), responsible AI innovation requires cooperation and the inclusion of diverse cultural perspectives. In this line, a cross-cultural study is being performed, involving Northern European, Southern European and East Asian (in particular Taiwanese) perspectives on RI with regards to the AI-based HAAL dashboard that is supposed to be a decision aid in home-based dementia care. Responsible AI innovation requires attention for individual situations and local context-of-use (Hagendorff, 2020), while at the same time catering to the need to offer somewhat standardized, scalable solutions (Peine & Moors, 2015). Therefore the insights from this study can provide researchers and AI practitioners with inspiration about how to address RI in and across contexts. [ABSTRACT FROM AUTHOR]

Purpose The world population is ageing with a decreasing workforce of care personnel, with dementia as one of the largest challenges in long-term care (Gauthier et al., 2021). Assistive technologies and services on several aspects of lifestyle could be of support because dementia impacts the daily lifestyle of people with dementia (PwD) and (in)formal caregivers. The HAAL project proposes an approach of an integrated ecosystem; a bundle of Ambient Assisted Living (AAL) products and an AI-driven dashboard. The vision of the HAAL project is to adopt Machine Learning to extract valuable information on a central dashboard. With global rapid development, user requirements for a state-of-the-art integrated ecosystem need iterative and local revisions. To achieve a feasible result, HAAL takes the Living Lab methodology (Bergvall-Kâreborn et al., 2009) as the core methodology - an approach of co-creation amongst stakeholders (i.e. researchers, PwD, informal caregivers, formal carers and technology developers). This research aims to study the end-users perceptions and opinions of an integrated-technology ecosystem. In addition, a rating and ranking study of the AAL products and technology combinations supports a careful consideration for implementation. Finally, the study sets the direction of the design of the future AI-driven dashboard. Method Three studies were conducted; user requirement investigation, HAAL technology demonstration and MoSCoW prioritisation (Kuhn, 2009). 32 PwD, 19 informal caregivers and 114 formal carers participated in the Netherlands, Italy, Taiwan and Denmark. The methods included interviews, focus groups, technological trials in demo rooms and survey research. Meaningful Tryout cards (Cornelissen & Suijkerbuijk, 2022) and working AAL products were used during the HAAL technology demonstration. The analysis techniques are thematic analysis and quantitative data exploration. User requirements were developed based on a motivational goal model (Burrows et al., 2018; Taveter et al., 2019). Results and Discussion PwD, informal caregivers and formal carers expressed the need to increase technology adoption to prolong the independent lives of PwD at home and avert the risk of health complications. In particular, formal carers recognise the potential of data utilisation. Four functional requirements for the ecosystem that emerged from the study are daily assistance, monitoring, alarm and prevention. According to end-users, the main emphasis should be on prevention (e.g., prompting actions before accidents happen and slowing dementia progression), because it could partially solve the current high workload and the shortages in the healthcare workforce. Fall detection and the GPS positioning are the best-rated features because of the concern of falling and wandering. The compulsory quality requirements of the ecosystem are usability and that the products are failure-proof. While giving back PwD autonomy was preferred in the Netherlands and Denmark, enriching the meaning of PwDs life was often addressed in Taiwan and Italy. The requirements serve as input for the iterative co-design of the AI-driven dashboard for formal carers. A potential challenge is related to change management. It is necessary to change the mindset of all end-users that digital care is care. [ABSTRACT FROM AUTHOR]