AI for Human Health: Developing Machine Learning Algorithms to Improve Personal Nutrition and Fitness; Supportive Coach or Military Officer? Understanding Designers and Users’ Analogies of Wearable Technology and Health Apps to Bridge the Gap in Mental Models

A startling contrast is evident in the observation that wearable technology such as smartwatches, similar health devices, and health apps continues to grow in popularity while physical fitness and nutrition perpetually decline. Addressing this issue, I propose integrating AI models with present devices to create individualized workout and nutrition plans that adapt to the user’s personal lifestyle and food habits. This implementation, however, must consider those who have mental health struggles and disorders associated with unhealthy mentalities. These mentalities then fuel behaviors encouraged by the existing software’s design. Thus, creating a product requires understanding how users perceive and interact with the design compared to what the designers intended. This insight helps identify the best way to encourage users to adopt healthier habits in a supportive, non-shaming manner.
The technical portion of my thesis produced a design concept of a smartphone or web application that is linked to a wearable device such as a watch or ring which measures nutrients absorbed into the body as well as monitors movement. The product would analyze input from the wearable to output recommended recipes based on lacking nutrients to increase their concentration through food consumption. Additionally, it would display a variety of fitness plans and schedules that account for the user’s preferences and agenda to reach goals in accordance with the desired timeline. For example, a user could inform the model that they love cardio workouts, particularly distance running, and they would like to train for a half-marathon to be ready in 6 months. Taking that data and connecting to the user’s calendar app, the AI would offer options to plan out times in the day to run with built in recovery periods based on their genetic profile and recent trends of requiring more or less recovery time after running certain distances.
In my STS research, I investigated the intentions of designers of health apps and wearable smart devices that monitor movement to compare them with a population of users dissatisfied with the product due to their experiences with it. My research achieved this by gathering discourse from both groups to compile it into an array of analogies for what type of person each felt best portrayed the applications. From this process, I gleaned the importance of developers engaging in discourse with users to think critically about what unintended consequences might be experienced so that features of the design might be adapted for the benefit of the user population. With little overlap between perception of a product among the two parties, unintended consequences are more likely to occur.
A sociotechnical perspective acknowledges that the technical affects the social, and therefore one’s actions in the technical realm holds consequences for society; indeed, people are affected by products. Those that build and design any sort of innovative technology must reflect upon what may result from its widespread release into the world. Engineering as social experimentation recognizes the partial ignorance that is involved in the design processes, followed by the continuous monitoring of results that allow iterative adjustments. With this as a core part of the system, engineers must engage in discourse with their audiences to comprehend each other’s mental models to ethically produce a technological artifact that successfully strives toward the greater good.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Briana Morrison

STS Advisor: Kathryn Neeley