Supporting dementia caregivers in Peru through chatbots: generative AI vs structured conversations
File(s)089_Espinoza_BCSHCI23.pdf (465.07 KB)
Published version
Author(s)
Espinoza Lau-Choleon, Fernanda
Cook, Darren
Butler, Chris
Calvo, Rafael
Type
Conference Paper
Abstract
In Peru, dementia caregivers face burnout, depression, stress, and financial strain. Addressing their needs involves tackling the intricacies of caregiving and managing emotional burdens. Chatbots can serve as a viable support mechanism in regions with limited resources. This study delves into the perceptions of dementia caregivers in Peru regarding a chatbot tailored to offer care navigation and
emotional support. We divided the study into three phases: the initial stage encompassed engaging stakeholders to define design requirements for the chatbot; the second stage focused on the creation of ‘Ana’, a chatbot for dementia caregivers; and the final stage assessed the chatbot through interviews and a caregiver satisfaction survey. ‘Ana’ was tested in two configurations - one
employed pre-defined conversation patterns, while the other harnessed generative AI for more dynamic responses. The findings reveal that caregivers seek immediate access to information on handling behavioural symptoms and a platform for emotional release. Moreover, participants
preferred the generative AI alternative of Ana, as it was perceived to be more empathic and human-like. The participants valued the generative approach despite knowing the potential risk of receiving inaccurate information.
emotional support. We divided the study into three phases: the initial stage encompassed engaging stakeholders to define design requirements for the chatbot; the second stage focused on the creation of ‘Ana’, a chatbot for dementia caregivers; and the final stage assessed the chatbot through interviews and a caregiver satisfaction survey. ‘Ana’ was tested in two configurations - one
employed pre-defined conversation patterns, while the other harnessed generative AI for more dynamic responses. The findings reveal that caregivers seek immediate access to information on handling behavioural symptoms and a platform for emotional release. Moreover, participants
preferred the generative AI alternative of Ana, as it was perceived to be more empathic and human-like. The participants valued the generative approach despite knowing the potential risk of receiving inaccurate information.
Date Issued
2023-08
Date Acceptance
2023-07-27
Citation
36th International BCS Human-Computer Interaction Conference (BCS HCI 23), 2023, pp.89-98
ISSN
1477-9358
Publisher
Association for Computing Machinery (ACM)
Start Page
89
End Page
98
Journal / Book Title
36th International BCS Human-Computer Interaction Conference (BCS HCI 23)
Copyright Statement
© Espinoza et al. Published by BCS Learning and Development Ltd. Proceedings of BCS HCI 2023, UK
License:
This work is licensed under a Creative Commons Attribution 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
License:
This work is licensed under a Creative Commons Attribution 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
License URL
Identifier
https://www.scienceopen.com/hosted-document?doi=10.14236/ewic/BCSHCI2023.11
Source
36th International BCS Human-Computer Interaction Conference 36th International BCS Human-Computer Interaction Conference Human-Computer Interaction Conference
Publication Status
Published
Start Date
2023-08-28
Finish Date
2023-08-29
Coverage Spatial
York, UK
Date Publish Online
2023-08