Approximately 70 percent of teens say that anxiety and depression are a major problem among their peers, and those concerns cross gender, racial, and socio-economic lines, according to the Pew Research Center.
Could robots help? Dr.
Elin Björling, co-founder of the Momentary Experience Lab at the University of Washington, leads a socio-therapeutic robot research program called Project EMAR, which helps teens moderate their moods and gain tools to deal with anxiety.
We spoke to Professor Björling about her research findings, which she'll be presenting at a robotics forum hosted by the National Science Foundation later this month.
Dr.
Bjorling, your lab has built the EMAR robot from the ground up.
It's designed to help teens chill out through sharing their feelings with the robot, getting hugs, and going through simple exercises.
Tell us how the project started.
I’ve been working with the teen population for two decades now and was noticing the rising levels in reported teen stress.
I decided to create a novel and engaging way to explore the subject.
During a conversation with Dr.
Emma Rose, now my Co PI [Principal Investigator] on Project EMAR, I jokingly suggested that we needed to design a robot as an engaging stress measurement and intervention tool for teens.
Although numerous smartphone apps have been made to help with stress, I’d become aware of how static—not very interactive, cluttered, and so many conversations at once—smartphones had become and wanted a more engaging, community-based research tool.
Robots seemed to be a good technology to explore for a research study, and that’s how it all started.
Elin Björling and EMAR
How much had already been done in the academic arena around teens, robots, and stress?
To be honest, I was absolutely shocked to find out how little research there was about social robots and teens, let alone projects involving teens in the design and development of a social robot.
Is it true that teens prefer telling the robot what's going on?
Yes, in our research, teens said they liked that EMAR cannot judge them or feel burdened by their problems.
Many teens have expressed to us that the fear of burdening others or causing worry is what keeps them from sharing with their peers of family members.
In your research paper, you also found that: 'Most teens are 1) highly optimistic about the helpfulness of robots, 2) do not feel nervous talking with a robot, but also 3) do not trust a robot with their data.' So how did you develop a human-robot trust model so the teens felt ready to share?
The first principle is transparency so our teen users can trust that EMAR is designed to help.
Making sure teens have agency over their data; giving them choice in their interactions related to privacy, or choosing to share them with peers, parents, or professionals, is an important principle.
There are, of course, strict rules about data capture/usage of minors so all data is anonymized and EMAR will never video record an interaction or record a person’s voice.
Likely data stored by EMAR will be aggregated stress levels, keywords, length of engagement, word count, and so on.
Rather than use an existing robot platform, you designed one from the ground up.
Yes, we started with ideation design sessions where teens sketched robot images and attributes, discussed roles and behaviors, and storyboarded school environment interactions.
From our initial ideation data, we began developing design principles and eventually requirements for our very first prototype.
For instance, teens quickly told us that elaborate, AI humanoids were NOT the type of robot they wanted to interact with at school.
They preferred boxy, non-humanoid form factors.
Tell us how you got the first prototype together.
Our first prototype (EMAR V1), based upon teen drawings and ideation, was a small, boxy, table-top robot that had only three questions and three different empathetic responses.
Right out of the gate teens were drawn to this robot, despite its very scrappy and clearly hand-made appearance.
From this we learned that "scrappy" might be working for us in terms of engagement.
From there we’ve iterated the design based upon design and interaction data with teens and now are just beginning EMAR V6.
The embodiment has changed over time.
How was the robot programmed? Is it ROS or a combination of other OS and open-source software programs, or completely custom/bespoke?
EMAR is a completely customized program, designed by Dr.
Maya Cakmak, our Co PI on the project.
She’s created a simple end-user visual programming tool similar to Google’s Blockly.
Because we have a design principle of transparency and customization, it’s important for our project that the final robot be completely customizable by its intended users: teens.
This means our "EMAR Stewards" (teen research participants), can take on EMAR customization and maintenance without any input from us, including EMAR’s data collection components to capture data that they feel their community is most interested in, such as sense of belonging or safety.
Describe the interaction process the current teens have with the robot to help them ameliorate their stress response to life.
We have many interactions we are designing with teens.
The primary ones involve sharing stress and mood levels with the robot through the belly touch screen and a disclosure interaction where teens vent their recent stressors verbally to the robot.
That whole self-reporting, or confessional, mode makes such sense for teens.
When I was a teenager I remember pouring my angst into a paper-based diary.
How else can they interact with the robot?
Our teen participants have a variety of other interactions, including "The Hug" in which teens can hug the robot.
Then, through haptic sensors, the EMAR delivers with a cute audio response and a visual pattern response with its LED lights.
Of course, there are many known physiological benefits to hugging, including the release of endorphins, the triggering of oxytocin, and the positive aspects of social connection through touch.
Soft toys, especially those with audio/visual response mechanisms, help small children manage sadness, but there are few tangible, or benign, outlets for teens.
EMAR could fit the bill.
Agreed.
In fact, the desire to hug EMAR is some of our strongest data over the years so that quickly became a technological requirement for the robot.
Tell us about EMAR’s other functionality.
We are just beginning to explore micro-interventions between EMAR and the teen participants.
Many of these are based on evidence-based interventions such as Dialectical Behavioral Therapy.
For example, EMAR can guide the teen through a short exercise to help them identify the emotions they are experiencing and provide a framework for them to share, explore, understand, and release the feeling, thereby helping reduce stress levels across the board.
What, if anything, has surprised you about the teens interaction with the robot to date?
I have worked with teens for two decades now, and I am continually surprised by their creativity and ideas.
They are the perfect collaborators for the co-design of new technologies with their heightened levels of sensitivity and empathy.
Much of the media today portrays teens as screen-addicted zombies incapable of empathy or sensitivity to others, but we now have four years of raw data suggesting that quite the opposite is true.
My takeaway in working with these teens is that they are in fact incredibly sensitive and empathetic beings who need an alternative outlet for their stress that frees them from worrying about others.
Recommended by Our Editors
Would you say the robot is both a social, and therapeutic, model? Is the distinction important, or not, in your view, necessary?
Yes, I would argue that ideally it is therapeutic because it is social.
Teens are social beings.
They are in a developmental stage where social relationships are incredibly important, so using a social robot as a tool for interaction seems developmentally appropriate.
We’ve also noticed that by inviting the teen to talk to the robot, teens in the environment seem to instinctively start talking to one another about their stress and experiences.
I am gently calling this an "invitation effect" of EMAR and we’ll have to study if this is true once we get to implementation.
EMAR actually increases human-to-human interaction?
Yes, in contrast to popular conceptions around social robots as "designed to replace a peer or a human," we've found quite the opposite is true.
Teens know EMAR is a tool to help them process their feelings and ideally make them feel a little bit better.
Their interactions with EMAR could increase their emotional literacy to be able to talk about their feelings with peers, parents and professionals.
What’s next in EMAR’s development process?
We are currently in NSF-funded year three and now heading into our final embodiment design that includes a moving neck, onboard haptic sensors, sounds, and lights.
After another few small studies to inform final requirements and ensure that our interaction designs are appropriate and engaging, we plan to run an implementation study in the fall.
Taking EMAR out of the lab?
Well, EMAR has never been in a lab so to speak.
All of our co-design sessions and research studies take place in local area high schools.
This helps to ensure that our data are contextually valid.
But yes, finally allowing EMAR to "live" on-site in a local area high school to find answers to the "big" questions, such as: "What impact does a social robot living in a school have on that community and that culture?" We will also attempt to determine EMAR’s usefulness as a measurement tool for teen stress levels, alongside its current functionality in helping relieve it.
These studies have been in schools, but still within a timed or somewhat controlled environment.
Eventually, do you see the robot being a permanent companion? As in Big Hero 6?
Teens LOVE Big Hero 6, and I think it’s because it represents an embodiment of empathy and that is what they are really craving.
Also much of our hug data from teens comes back to BH6.
Ideally EMAR is a community-based tool, customized by, and serving, the community in which it lives.
Although in the future, I’m sure we’ll want to explore a home version of EMAR for a specific population of adolescents, such as those struggling with depression, and other clinical conditions or populations.
We believe there are many possibilities for where EMAR might be helpful in the future.
Finally, how did you personally get into this field, and what made you decide to pursue your PhD in Nursing Science?
I started college as a dual dance and film major.
Through dance I really became interested in the mind-body connection, which drew me to health psychology.
After studying health psychology I became interested in chronic illnesses related to stress, so undertook an interdisciplinary health sciences degree at UW to learn how to conduct research with human populations.
It was there that I became fascinated with adolescent health and, crucially, the importance of the adolescent developmental stage.
It’s an amazing but such fragile time, and mental health-based interventions are both necessary and powerful.
