Artificial Intelligence Chatbot Behavior Change Model for Designing Artificial Intelligence Chatbots to Promote Physical Activity and a Healthy Diet: Viewpoint

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The search returned 108 articles from four databases, with 15 published in 2020, 26 in 2019, 15 in 2018, 14 in 2017, five in 2016, and the remaining 33 from 2015 or before.

“In total, the search returned 108 articles from the four databases, with 15 published articles in 2020, 26 in 2019, 15 in 2018, 14 in 2017, five in 2016, and the remaining 33 from 2015 or before.”

May combine two claims.

Four out of seven (57.1%) studies reported chatbots as the only intervention used to deliver behavior change strategies, whereas the other three articles reported chatbots as an auxiliary component complementing other intervention approaches.

“It is worth noting that four out of seven (57.1%) studies reported chatbots as the only intervention used to deliver behavior change strategies [26,43,44,46], whereas the other three articles reported chatbots as an auxiliary component complementing other intervention approaches such as messages and conversations delivered by human facilitators [41,42,45] (Multimedia Appendix 1).”

May combine two claims.

None of the reviewed studies reported in detail how they developed the chatbot program and none discussed ethical considerations regarding issues such as transparency, privacy, and potential algorithmic biases.

“None of the studies reported in detail how they developed the chatbot program and none discussed ethical considerations regarding issues such as transparency, privacy, and potential algorithmic biases.”

Probably not a substantial claim: may not pass S3 (load-bearing).

After screening, 101 (93.5%) articles were excluded for reasons including commentary or opinion pieces, scoping reviews, or empirical studies that addressed health domains other than physical activity and diet.

“After the screening, 101 (93.5%) articles were excluded for the following reasons: commentary or opinion pieces, scoping reviews, or empirical studies that addressed health domains other than physical activity and diet (eg, chatbots assisting diagnostic tasks or offering mental health interventions or treatment).”

May combine two claims.

The reviewed chatbots were designed with different theoretical components and varied in their abilities to engage in natural language conversations, relationship building, and emotional understanding.

“The reviewed chatbots were designed with different theoretical components and varied in their abilities to engage in natural language conversations, relationship building, and emotional understanding.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Physical inactivity and an unhealthy diet are leading risk factors for noncommunicable diseases and death worldwide.

“Physical inactivity and an unhealthy diet continue to be some of the leading risk factors for noncommunicable diseases (NCDs), such as cardiovascular disease, diabetes, and obesity [1,2], and death worldwide [3].”

• cites [1] sourceNot checked yet

  Not checked in the free run.

• cites [2] sourceNot checked yet

  Not checked in the free run.

• cites [3] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

BERT and GPT2 produce outputs that are difficult for humans to interpret and can make errors that violate common senses in specific domains because they operate on representation space and do not have access to symbolic common-sense information.

“As these models operate on representation space and do not have access to symbolic common-sense information, they produce outputs that are difficult for humans to interpret and can make errors that violate common senses in specific domains.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Chatbot technologies have the potential to understand individuals through natural human conversations, persuade individuals to change, and build sustaining supportive relationships for maintaining healthy behaviors.

“In other words, chatbot technologies have the potential to “understand” individuals through natural human conversations, persuade individuals to change, and build sustaining supportive relationships for maintaining healthy behaviors.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A large number of published studies support the effectiveness of using digital technologies to deliver intervention materials to diverse populations.

“In the past two decades, there has been a large number of published studies using internet and mobile-based behavior interventions to support the effectiveness of using digital technologies to deliver intervention materials to diverse populations [8,14].”

• cites [8] sourceNot checked yet

  Not checked in the free run.

• cites [14] IJsselsteijn, W, 2006, Persuasive Technology. PERSUASIVE 2006. Lecture Notes in Computer ScienceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Several papers have shown that pretrained models can be tailored for task-oriented dialog generation, such as for conversations about restaurant recommendations and donation persuasion.

“For instance, several papers have shown that pretrained models can be tailored for task-oriented dialog generation, such as for conversations about restaurant recommendations and donation persuasion [39,40].”

• cites [39] WuQZhangYLiYYuZAlternating roles dialog model with large-scale pre-trained language modelsarXiv20192020-08-29https://arxiv.org/abs/1910.03756Not checked yet

  Not checked in the free run.

• cites [40] GuJWuQWuCShiWYuZA tailored pre-training model for task-oriented dialog generationarXiv20202020-08-28https://arxiv.org/pdf/2004.13835.pdfNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

One relevant review pointed out that the interpretation and application of behavior change theories were usually not reported.

“One relevant review [33] focused on discussing the development of embodied conversational agents for a healthy lifestyle, and pointed out that the interpretation and application of behavior change theories were usually not reported.”

• cites [33] sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Recent success of large pretrained language models such as BERT and GPT2 provides promising opportunities to incorporate language priors to downstream natural language processing tasks.

“The recent success of large pretrained language models, such as Bidirectional Encoder Representations from Transformers (BERT) developed by Google [37] and Generative Pre-Training-2 (GPT2) developed by Open AI [38], provides promising opportunities to incorporate language priors to down-stream natural language processing (NLP) tasks.”

• cites [37] DevlinJChangMLeeKToutanovaKBERT: Pre-training of Deep Bidirectional Transformers for Language UnderstandingProceedings of NAACL-HLT 20192019The 17th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT 2019)June 7, 2019Minneapolis, MinnesotaNot checked yet

  Not checked in the free run.

• cites [38] RadfordAWuJChildRLuanDAmodeiDSutskeverILanguage models are unsupervised multitask learnersOpenAI Blog20192020-08-29https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdfNot checked yet

  Not checked in the free run.

May combine two claims.

It remains unclear how to evaluate a chatbot's efficacy, the theoretical mechanisms through which chatbot conversations influence users, and potential ethical problems.

“Consequently, it remains unclear how to evaluate a chatbot’s efficacy, the theoretical mechanisms through which chatbot conversations influence users, and potential ethical problems.”

May combine two claims.

In a feasibility study, adolescent patients actively engaged with the Tess behavioral coaching chatbot, reported experiencing positive progress toward their goals, and deemed the chatbot helpful.

“Patients actively engaged with the chatbot, reported experiencing positive progress toward their goals, and deemed the chatbot helpful.”

• cites [44] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The computers are social actors (CASA) paradigm assumes that humans can develop positive social relations with computer systems as the human familiarity of the system increases.

“While the CASA paradigm assumes that humans can develop positive social relations with computer systems as the human familiarity of the system increases, the UVE argues that too much human familiarity would bring feelings of eeriness and discomfort.”

• cites [57] Reeves, B, 1996, The Media Equation: How People Treat Computers, Television, and New Media Like Real People and PlacesNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

The AI chatbot behavior change model includes four major components: designing chatbot characteristics and understanding user backgrounds; building relational capacity; building persuasive conversational capacity; and evaluating mechanisms and outcomes.

“We named this framework the AI chatbot behavior change model, which includes the following four major components: (1) designing chatbot characteristics and understanding user backgrounds; (2) building relational capacity; (3) building persuasive conversational capacity; and (4) evaluating mechanisms and outcomes.”

• AssumptionThe authors state this as an assumption: A personalized intervention is more effective than a nonpersonalized one because it tailors both behavior change strategies and persuasive messaging to each user’s unique background and needs to achieve personally optimized outcomes.

May combine two claims.

Thirty percent of participants stopped using the app over the course of the study, presenting a challenge for the chatbot's ability to engage participants.

“However, 30% of participants stopped using the app over the course of the study, presenting a challenge for the chatbot’s ability to engage participants.”

• cites [42] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

In a proof-of-concept study of the Paola chatbot, participants reported relevant weight loss and improved diet.

“The results showed that participants reported relevant weight loss and improved diet.”

• cites [45] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The Reflection Companion conversations successfully triggered self-reflections that led to increased motivation, empowerment, and adoption of physical activity behaviors.

“The conversations successfully triggered self-reflections that led to increased motivation, empowerment, and adoption of physical activity behaviors (eg, walking to a grocery store instead of taking a car) [26].”

• cites [26] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Behavior change chatbots are a type of persuasive technology, which is more complicated than designing a social chatbot to engage in general conversations.

“We conceptualize behavior change chatbots as a type of persuasive technology [14], which is more complicated than designing a social chatbot to engage in general conversations (eg, talking about movies or weather) [47].”

Classified as probably not a substantial claim.

The uncanny valley effect (UVE) argues that too much human familiarity would bring feelings of eeriness and discomfort.

“While the CASA paradigm assumes that humans can develop positive social relations with computer systems as the human familiarity of the system increases, the UVE argues that too much human familiarity would bring feelings of eeriness and discomfort.”

• cites [58] sourceNot checked yet

  Not checked in the free run.

• cites [59] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The Ally chatbot system by Kramer et al was able to welcome each participant using personalized messages and track individual physical activity using the smartphone's built-in accelerometer.

“In the realm of physical activity chatbot interventions, the Ally chatbot system by Kramer et al was able to welcome each participant using personalized messages and track individual physical activity using the smartphone’s built-in accelerometer [42].”

• cites [42] sourceNot checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: A personalized intervention is more effective than a nonpersonalized one because it tailors both behavior change strategies and persuasive messaging to each user’s unique background and needs to achieve personally optimized outcomes.

May combine two claims.

Some studies show evidence that people respond well and disclose more personal information if the chatbot is presented as a bot and can also display emotions.

“However, contrary findings have also been identified as some studies show evidence that people respond well and disclose more personal information if the chatbot is presented as a bot and can also display emotions [60,61].”

• cites [60] sourceNot checked yet

  Not checked in the free run.

• cites [61] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The Ally system specifically set a personalized activity goal slightly above the participant's current average activity level.

“The system specifically set a personalized activity goal slightly above the participant’s current average activity level.”

• cites [42] sourceNot checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: A personalized intervention is more effective than a nonpersonalized one because it tailors both behavior change strategies and persuasive messaging to each user’s unique background and needs to achieve personally optimized outcomes.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Commonly used behavior change theories include motivational interviewing, the social cognitive theory, the transtheoretical model, and the theory of planned behavior.

“Commonly used behavior change theories include motivational interviewing [81], the social cognitive theory [56], the transtheoretical model [82], and the theory of planned behavior [83].”

• cites [81] Miller, W, 2012, Motivational interviewing: helping people changeNot checked yet

  Not checked in the free run.

• cites [56] sourceNot checked yet

  Not checked in the free run.

• cites [82] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

One study reported that users of the Replika companion chatbot perceived the chatbot to be human-like, intelligent, supportive, and able to facilitate social connection; however, uncanny valley effects also emerged as some users felt the chatbot's conversation was too natural and thus "creepy."

“One study reported that users of a companion chatbot (called “Replika”) perceived the chatbot to be human-like, intelligent, supportive, and able to facilitate social connection. However, UVEs also emerged as some users felt that the chatbot’s conversation was too natural and thus “creepy” [73].”

• cites [73] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

A case study analyzing user reviews of the Amazon chatbot device found that over half of the reviewers referred to the chatbot using the personified name "Alexa," and as users' social interactions with the device increased, a greater level of personification occurred, which was associated with increased product satisfaction.

“In another case study that analyzed user reviews of the Amazon chatbot device, researchers found that over half of the reviewers referred to the chatbot using the personified name “Alexa,” and as users’ social interactions with the device increased, a greater level of personification occurred, which was associated with increased product satisfaction [74].”

• cites [74] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

One common limitation of traditional programs is the static nature of persuasive messages because of infrequent measurements of behaviors and users' behavior change stages.

“One common limitation of traditional programs is the static nature of persuasive messages, because of infrequent measurements of behaviors and users’ behavior change stages.”

Classified as probably not a substantial claim.

Bickmore et al provided extensive discussions on the principles of building relational capacity in behavior change agents, such as using social dialog, empathy, meta-relational communication, humor, self-disclosure, persistent memory, and agent variability.

“Bickmore et al provided extensive discussions on the principles of building relational capacity in behavior change agents, such as using social dialog, empathy, meta-relational communication (talk about the relationship), humor, self-disclosure, persistent memory, and agent variability [70].”

• cites [70] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

People tend to personify Alexa, which is in line with the CASA paradigm.

“This suggests that people tend to personify Alexa, which is in line with the CASA paradigm.”

May combine two claims.

One study showed that when compared to a nonrelational agent, a relational agent was more respected, liked, and trusted, which led to more positive behavior changes.

“One of their studies showed that when compared to a nonrelational agent, a relational agent was more respected, liked, and trusted, which led to more positive behavior changes [29].”

• cites [29] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Although the question-behavior effect was small and based on survey reports, it suggests that questions can function as a reminder or cue to action.

“Although this effect was small and based on survey reports, it suggests that questions can function as a reminder or cue to action.”

May combine two claims.

Research has shown that an accelerometer installed on smartphones is accurate for tracking step count and that GPS signals can be used to estimate activity levels.

“For instance, research has shown that an accelerometer installed on smartphones is accurate for tracking step count [9] and that GPS signals can be used to estimate activity levels [87].”

• cites [9] sourceNot checked yet

  Not checked in the free run.

• cites [87] sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

The proposed AI chatbot behavior change model considers a wide range of chatbot-related components (chatbot/user characteristics, relational capacity, persuasive conversational capacity) and points out potential mediating and moderating factors to be evaluated to establish the efficacy of chatbots in changing physical activity and diet behaviors, as well as health outcomes.

“The strength of the proposed model is that it considers a wide range of chatbot-related components, including chatbot/user characteristics, relational capacity, and persuasive conversational capacity, and points out potential mediating and moderating factors to be evaluated to establish the efficacy of chatbots in changing physical activity and diet behaviors, as well as health outcomes.”

Probably not a substantial claim: may not pass S4 (based).

This is the first theoretical framework to provide a guideline to design and evaluate chatbot-based physical activity and diet behavior interventions.

“To our knowledge, this is the first theoretical framework to provide a guideline to design and evaluate chatbot-based physical activity and diet behavior interventions.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

These intermediate changes are important to explain the mechanisms of chatbot interventions and to design more effective interventions in the future.

“These intermediate changes are important to explain the mechanisms of chatbot interventions and to design more effective interventions in the future.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

The proposed theoretical framework is the first step to conceptualize the scope of the work and to synthesize all possible dimensions of chatbot features to inform intervention design.

“Our proposed theoretical framework is the first step to conceptualize the scope of the work and to synthesize all possible dimensions of chatbot features to inform intervention design.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Chatbots can lead people to change their perceptions of themselves (eg, attitude, self-efficacy, and perceived social support) and help people to shape and form new behavior choices and patterns.

“Chatbots can lead people to change their perceptions of themselves (eg, attitude, self-efficacy, and perceived social support) and help people to shape and form new behavior choices and patterns.”

May combine two claims.

These results indicate that AI chatbots can adapt not only behavior change goals and techniques, but also conversational styles based on learning from a user's natural language inputs to enhance the engagement and effectiveness of messages.

“These results indicate that AI chatbots can adapt not only behavior change goals and techniques, but also conversational styles (eg, emotional tones) based on learning from a user’s natural language inputs to enhance the engagement and effectiveness of messages.”