JMIR Publications

ABSTRACT

Background:

Overweight and obesity have now reached a state of a pandemic. Existing clinical and commercial programs are inefficient in enabling and sustaining weight loss. Artificial intelligence (AI) chatbots have strong potential in supplementing long-term health coaching for efficient and effective weight loss.

Objective:

To provide an overview of the potential uses of AI-chatbots for weight loss in people with overweight and obesity and identify the essential components to prolong user engagement.

Methods:

A scoping review was conducted using the Arksey and O’Malley’s 5-stage framework. Articles were searched across nine electronic databases (ACM Digital library, CINAHL, Cochrane–Central, Embase, IEEE Xplore, PsycINFO, PubMed, Scopus, and Web of Science) from inception till 9 July 2021. Grey literature, reference lists, and google scholar were also searched.

Results:

23 studies representing 2,231 participants were included and evaluated in this review. Majority of the studies focused on using AI-chatbots to promote a healthy diet and exercise (n=8), three studies used AI-chatbots solely for lifestyle data collection and obesity risk assessment while only one was on promoting a healthy diet, exercise and stress management. 43.5% of the studies used solely text-based AI-chatbots, 56.3% operationalized the AI-chatbots through smartphones and 39.1% integrated data collected through fitness wearables or internet-of-things appliances (e.g. smart refrigerators to track food consumption). Core functions of AI-chatbots were providing personalized recommendations (n=20; 87.0%), motivational messages (n=18; 78.3%), gamification (n=6; 26.1%) and emotional support (26.1%). Study participants who experienced speech- and augmented reality-based in addition to text-based chatbot interactions had reportedly higher user engagement due to the convenience of hands-free interactions. Enablement of conversations through multiple platforms (e.g. SMS text messaging, Slack, Telegram, Signal, WhatsApp or Facebook Messenger) and devices (e.g. laptops, Google Home and Amazon Alexa) were reported to increase user engagement. The human semblance of a chatbot through verbal and non-verbal cues improved human-chatbot rapport and user engagement through interactivity, relatability and empathy. Other techniques used in text-based chatbots were the use of personally- and culturally-apt colloquial tones and content, emojis to emulate human emotional expressions, positively-framed words, citations of credible information sources, chatbot personification, validation and the provision of real-time, fast and reliable recommendations. Prevailing issues included privacy, accountability, user burden during chatbot engagement and the interoperability with other databases, third-party apps, social media platforms, devices and appliances.

Conclusions:

AI-chatbots should be designed to be human-like, personalized, contextualized, immersive and enjoyable to enhance user experience, engagement, behavior change and weight loss. These require the integration of health metrics (e.g. based on self-reports and wearable trackers), personality and preferences (e.g. based on goal-achievements), circumstantial behaviors (e.g. trigger-based overconsumption) and emotional states (e.g. chatbot conversations and wearable stress detectors) to deliver personalized and effective recommendations for weight loss. Clinical Trial: NIL