Insights and lessons learned from trialling a mental health chatbot in the wild

Potts, Courtney; Bond, Raymond; Mulvenna, Maurice; Ennis, Edel; Bickerdike, Andrea; Coughlan, Edward K.; Broderick, Thomas; Burns, Con; McTear, Michael F.; Kuosmanen, Lauri; Nieminen, Heidi; Boyd, Kyle; Cahill, Brian; Vakaloudis, Alex; Dhanapala, Indika; Vartiainen, Anna-Kaisa; Kostenius, Catrine; Malcolm, Martin

doi:10.1109/iscc53001.2021.9631395

Cited by 2 publications

(9 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One thread for improving engagement has been to embed chat agents into embodied characters, for example, a virtual character for delivering cognitive behavior therapy or related approaches [ 8 , 35 , 36 , 41 ]; this embodiment may elicit more engagement than a purely text interface [ 8 ]. Social robots have similarly been used for cognitive behavior therapy [ 42 ] and positive psychology [ 20 , 37 ], or more generally designed to be companions to support well-being [ 7 , 20 , 43 ]. While some projects indicate positive well-being outcomes (e.g.…”

Section: Related Workmentioning

confidence: 99%

“…While some projects indicate positive well-being outcomes (e.g. [ 8 , 35 , 42 ]), overall results are mixed with reviews failing to find consistent benefits [ 34 , 38 ], and user drop-out rates remain high [ 20 ].…”

Section: Related Workmentioning

confidence: 99%

“…More autonomous solutions still often require structured environments with expert supervision (e.g. [ 20 , 45 ]), precluding domestic deployment. Such setups can diminish the validity of results (e.g., as with the Hawthorne effect [ 47 ]), as participants are interacting in highly controlled unnatural (e.g.…”

Section: Related Workmentioning

confidence: 99%

“…Autonomous natural dialog interaction remains elusive for interactive agents; both stable speech recognition and semantic parsing remain an ongoing challenge [ 20 , 48 ]. As a result, a typical format is agent-led content delivery, where an agent monitors for expected utterances, keywords, or events and provides pre-scripted responses and dialog flow (e.g., [ 35 , 37 ]).…”

Section: Related Workmentioning

confidence: 99%

“…However, successful self-reflection agent deployment remains elusive for a range of reasons (e.g., see [ 19 ]) including the technical challenges of speech recognition and parsing [ 20 ] and overarching privacy concerns [ 21 ]. Within this context, we aimed to bypass many of these problems by removing the requirement for the robot to parse the content of user utterances – we propose that a conversational behavior design can provide many of the self-reflection benefits of conversational agents, even without recording or parsing what a person says.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Say what you want, I’m not listening!

González

Geiskkovitch

Young

2023

I-Com

View full text Add to dashboard Cite

We present a conversational social robot behaviour design that draws from psychotherapy research to support individual self-reflection and wellbeing, without requiring the robot to parse or otherwise understand what the user is saying. This simplicity focused approached enabled us to intersect the well-being aims with privacy and simplicity, while achieving high robustness. We implemented a fully autonomous and standalone (not network enabled) prototype and conducted a proof-of-concept study as an initial step to test the feasibility of our behaviour design: whether people would successfully engage with our simple behaviour and could interact meaningfully with it. We deployed our robot unsupervised for 48 h into the homes of 14 participants. All participants engaged with self-reflection with the robot without reporting any interaction challenges or technical issues. This supports the feasibility of our specific behaviour design, as well as the general viability of our non-parsing simplicity approach to conversation, which we believe to be an exciting avenue for further exploration. Our results thus pave the way for further exploring how conversational behaviour designs like ours may support people living with loneliness.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Say what you want, I’m not listening!

González

Geiskkovitch

Young

2023

I-Com

View full text Add to dashboard Cite

show abstract

A Mental Health and Well-Being Chatbot: User Event Log Analysis

Booth¹,

Potts²,

Bond³

et al. 2023

JMIR Mhealth Uhealth

Self Cite

View full text Add to dashboard Cite

Background Conversational user interfaces, or chatbots, are becoming more popular in the realm of digital health and well-being. While many studies focus on measuring the cause or effect of a digital intervention on people’s health and well-being (outcomes), there is a need to understand how users really engage and use a digital intervention in the real world. Objective In this study, we examine the user logs of a mental well-being chatbot called ChatPal, which is based on the concept of positive psychology. The aim of this research is to analyze the log data from the chatbot to provide insight into usage patterns, the different types of users using clustering, and associations between the usage of the app’s features. Methods Log data from ChatPal was analyzed to explore usage. A number of user characteristics including user tenure, unique days, mood logs recorded, conversations accessed, and total number of interactions were used with k-means clustering to identify user archetypes. Association rule mining was used to explore links between conversations. Results ChatPal log data revealed 579 individuals older than 18 years used the app with most users being female (n=387, 67%). User interactions peaked around breakfast, lunchtime, and early evening. Clustering revealed 3 groups including “abandoning users” (n=473), “sporadic users” (n=93), and “frequent transient users” (n=13). Each cluster had distinct usage characteristics, and the features were significantly different (P<.001) across each group. While all conversations within the chatbot were accessed at least once by users, the “treat yourself like a friend” conversation was the most popular, which was accessed by 29% (n=168) of users. However, only 11.7% (n=68) of users repeated this exercise more than once. Analysis of transitions between conversations revealed strong links between “treat yourself like a friend,” “soothing touch,” and “thoughts diary” among others. Association rule mining confirmed these 3 conversations as having the strongest linkages and suggested other associations between the co-use of chatbot features. Conclusions This study has provided insight into the types of people using the ChatPal chatbot, patterns of use, and associations between the usage of the app’s features, which can be used to further develop the app by considering the features most accessed by users.

show abstract

Insights and lessons learned from trialling a mental health chatbot in the wild

Cited by 2 publications

References 23 publications

Say what you want, I’m not listening!

Say what you want, I’m not listening!

A Mental Health and Well-Being Chatbot: User Event Log Analysis

Contact Info

Product

Resources

About