Overcoming bias in Voice UI

How a workshop uncovered what we thought we already knew.

Chantal Jahchan
Jun 15, 2017 · 6 min read

Each summer Moment creates a research project about an exciting topic we see on the horizon. In the past, our teams have researched self-driving cars, way-finding with Google Glass, consumer-facing media consumption, symptom and medication management for cancer patients, and in 2016, Virtual Reality in the classroom with Peer.

For Moment’s 2017 summer project, our group of designers, interns, leaders, and experts will explore the intersection of Voice User Interfaces (VUI) and first-time parents. We’ll be thinking about how a VUI tool could guide new parents through their child’s cognitive development by adapting to the child’s unique progress and needs.

Preparing to design for Voice UI

As designers, before we even begin to understand the Voice User Interface (VUI) playing field — think Siri, Alexa, Google Home — it’s important to get our biases about this relatively new technology into the open.

Despite rapidly increasing sales and media buzz, because we’ll be working with a type of product that has yet to prove its endurance and feasibility in the marketplace, we must ensure that our early decisions don’t rest on assumptions, preconceptions, or, worse yet, misconceptions.

“The only true wisdom is in knowing you know nothing.”

Discussing our limited exposure to Amazon Alexa.

We decided that conducting a bias workshop would be a great first step in preparing to design for Voice UI. We sat together in a room and did several quick-minute sprints to write out our biases onto sticky notes around several subjects as they relate to VUI. Then each of us took turns explaining our biases to the group.

Through the process of identifying and discussing our preconceived biases — whether rooted in skepticism or optimism — we hoped to prepare ourselves to ideate openly and design with discretion.

Insights from the workshop

We now know what we thought we knew about VUI. In other words, we narrowed the gap between what we thought we knew and what we actually knew (see below).

By the time we begin to design — after we’ve conducted primary and secondary research — we hope to expand our knowledge and understanding of VUI (that’s what we mean by “eventually” in the diagram below).

We learned a lot about ourselves as users and observers of VUI devices and synthesized our findings into six key principles, along with direct quotes from our group session.

1. We hesitate to use VUIs because of social stigmas and fears.

Illustration by Andy Rementer for The New Yorker.

2. We’re put off by the speech and tone assigned to VUI personas.

A list of things a user can ask Siri in iOS 9. Source: theappfactor.com

3. We think we know the limitations of voice-first interfaces.

  • “VUI is only useful for binary questions and answers.”
  • “VUI requires users to remember a long list of things they can ask for.”
  • VUI is only useful in a handful of unique situations.”
  • “What can Siri/Alexa do that I can’t already do from my iPhone?”
  • “Bad VUI is far worse than bad GUI.”
  • “VUI doesn’t work well on its own and needs support from GUI.”
  • “VUI causes more user frustration than GUI does.”
  • “I am a visual person and prefer text/photos to audio/voice.”

4. We think we know the short list of the things VUI is good for.

  • Ease of tasks: “VUI makes tedious tasks simple to do.”
  • “VUI flattens navigation structures for more efficiency.”
  • Typing is slower than speaking.”
  • “VUI is useful during activities like driving or cooking where my hands and eyes are occupied.”
Come on, Siri. It’s 2017. Source: searchenginewatch.com

5. We think tech limitations stand in the way of VUI’s improvement.

  • “I expected VUI technology to be more advanced by now.”
  • “I wouldn’t want to interrupt my music to speak or listen to Alexa.”
  • VUI is useless without internet access: “Alexa can’t even turn off my lights when the internet fails.”
  • “VUI is buggy.”
  • Cloud storage: “What if I lose my data?”
Jason confronts his data-sharing fears, one Post-it® at a time.

6. We don’t trust the intentions of VUI companies when it comes to the privacy and use of our personal data.

Where to next?

Before we begin designing a new VUI product for first-time parents, our group will conduct primary research with actual new, first-time parents. Through in-depth interviews, we look to understand the experience of parenthood from those who’ve experienced it.

We’ll identify the resources parents use to fill knowledge gaps, take note of rituals, games, and activities they use to promote their child’s cognitive development, and understand the distribution of responsibilities between primary and secondary caregivers.

Now that we have our biases out in the open, we’ll be able to tackle some of them head-on. Stay tuned to our progress throughout the summer!

Thomas Carroll, Blue Cuevas, Chantal Jahchan, and Jason Kim are interns at Moment in New York. Thomas is pursuing an MDes in Experience Design at VCU Brandcenter, Blue is pursuing an MDes in Design Strategy at IIT Institute of Design, Chantal is an incoming senior at Washington University in St. Louis pursuing a BFA in Communication Design, and Jason Kim is a recent graduate of Rensselaer Polytechnic Institute with a BS in Interdisciplinary Design. They’re currently exploring the intersection of voice user interface and first-time parents. You can follow the team’s progress this summer on Momentary Exploration.

Design Intelligence

Digital undercurrents, the evolution of industry, and the…