Even with fitness bands and smartwatches fully ready to monitor and report our activity, we don’t stick with our resolutions to get fit and stay fit. Why? Because these devices aren’t sufficiently engaging and make some people feel inadequate if they don’t reach their goals, says James Landay, professor of computer science at Stanford University and associate director of the Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI). “They become a reminder of what we’re not doing that we committed to do.”
To address this problem, Landay and a crew of students created WhoIsZuki, a fitness app that uses storytelling to keep users interested in staying active over time. Each week, if the user meets his or her goals, the story moves to the next chapter in the ongoing tale of Zuki, an adorable extraterrestrial who travels to Earth to find his lost brother. The story is told visually, with only minimal text, and appears on the home screen of the user’s smartphone.
“Just by seeing the display every time you unlock your phone, we hope you will be encouraged to do more physical activity,” Landay says.
In a recent paper, Landay’s team described both the iterative steps they took to make Zuki’s story as engaging as possible and the pilot study they ran to test the app. Although the pilot involved only 16 test subjects over the course of three weeks, the team found that the users who could earn new chapters each week stayed more engaged than the control group for whom the app’s narrative repeats each week. The publication won best paper at the Association for Computing Machinery’s Conference on Human Factors in Computing Systems (CHI 2020). The team is now wrapping up a longer study with 40 subjects over four and a half months.
Landay himself is a committed user of WhoIsZuki. When Saturday night rolls around, if he hasn’t met his goals for the week, he’ll often go out for a walk or bike ride just so he can make it to the next chapter. “It definitely pushes you over the bar.”
Value of Stories
The value of narrative as a motivator only became clear to Landay a decade ago when he was working on a sustainability app called UbiGreen. The team was testing different ways to reward people for engaging in green activities. In one version, users earned leaves on a tree, which reset each week with an empty tree. In another version, a polar bear sat on a small iceberg, and as the user engaged in green activities, Landay says, “the ice grew, and then the polar bear got a mate, and then there were more fish and seabirds, and they had cubs, and eventually the northern lights came out.” Users found these visual cues motivating: They took a bus or rode a bike to work so they could get to the next part of the “game” or “story.” But Landay’s team hadn’t designed a game or story. “We were just using a visual way of communicating the information, but people were seeing story and narrative in just that simple visualization.”
This led to the idea of motivating behavior change by designing an engaging narrative with interesting characters and a plot that thickens and builds tension to a climax.
But the WhoIsZuki backstory doesn’t end there. The team sought help from narratologist and Stanford English professor Paula Moya. The initial storyline, which involved Zuki collecting plants and animals from Earth to take back to his home planet, didn’t resonate with users. Moya advised making it more personal, which led to the idea of Zuki being on a mission to find his brother. Users also preferred chapters where Zuki was climbing, jumping or swimming, Landay says, “because those were more connected with their own behavior change efforts.” And they didn’t want much text, which meant the story had to be made visually clear at a glance – as well as compelling. Using Moya’s expertise, the team fine-tuned the less active chapters. Users also cared about small design decisions – such as having progress indicators that were tied to the story (e.g., carabiners for the mountain climbing segment). “We were surprised that these little design details mattered,” Landay says.
If the team’s current 40-person study pans out, Landay wants to compare WhoIsZuki’s engagement level with a version where users perform both physical and green activities to advance to the next chapter, a version with culturally appropriate stories for different audiences, or one that helps users set goals to improve their exercise over time rather than just being consistent each week.
In the long run, any commercial version of WhoIsZuki will need to have a menu of stories for users to choose from. “We’re definitely of the mind that there will be different stories that will appeal to different audiences,” Landay says.
Will narrative be enough to keep users engaged in exercise long term? That’s hard to say. Ultimately, it will be about users’ mindset: Can they learn to think of exercise as something fun and enjoyable rather than a slog? We’ll have to wait until the end of this story to find out.
Using Narrative and AI for Education: Smart Primer
Landay and his colleagues are also working on a project called Smart Primer that uses narrative on a tablet-like device to engage with 7- to 11-year-old students. The idea for Smart Primer sprang from a science fiction novel by Neal Stephenson called The Diamond Age: Or, A Young Lady’s Illustrated Primer. In the book, a 4-year-old girl gets hold of a futuristic tablet computer that uses both visual and textual narrative to teach her – over the course of her childhood and well into her teens – not only how to read and write and do math problems, but also survival skills such as martial arts. “In the book, although all of the learning activities are embedded in the narrative,” Landay says, “some are on the tablet, while others require the girl to do things in the world.”
Based on Stephenson’s science fiction premise, Landay and his team are developing the Smart Primer, a device that uses narrative and AI to serve up educational material that fits a student’s needs. “With the Smart Primer,” Landay says, “we’re using narrative and AI as a way to improve engagement – in this case, with a learning activity instead of physical exercise.”
Smart Primer prototypes have gone through several iterations during their first four years of development. Most recently, the team has been testing a chatbot that will help children when they get stuck and don’t know what to do. It will take some innovative AI work to figure out the level of difficulty a student can handle as well as what help or hints to provide, Landay says. But in a recently published research paper, he and his team showed that children who used a Smart Primer that offered both narrative and chatbot features learned more and were more engaged than students who used the device without those features – and they also retained the information longer.
Going forward, Landay hopes to have students using the Smart Primer to do things in the real world – “like Pokémon Go for education,” he says.
— Katharine Miller