FCB and Google Strive to Make Voice Assistants Work for People With Down Syndrome

Imagine asking your voice platform of choice to turn on the lights, play some tunes or provide a weather report—and the system refusing to comply because it can’t comprehend your speech patterns. 

Folks with Down syndrome face such frustrations, with Google reporting that its technology misunderstands about one-third of the words spoken by those living with the developmental disorder.

To devise a solution, people with Down syndrome are being asked to record various phrases online. Those sound files will be added to Google’s algorithm to improve its grasp of vocal idiosyncrasies caused by variations in facial skeletal and muscular structures. Dubbed “Project Understood,” the initiative launched Monday during Canadian Down Syndrome Week. It marks the latest innovative collaboration from Canadian Down Syndrome Society and FCB Canada. 

The video below features folks with Down syndrome demonstrating the great things they can do with their voices—except make Google’s voice assistant understand requests and commands. 

This next clip illustrates the team’s approach to enhancing Google’s algorithm:

According to Jupiter Research, there will be 8 billion voice assistants in the world by 2023. They’ll perform increasingly essential functions, so expanding the machines’ range of understanding has become a priority.

“For most people, voice technology simply makes life a little easier,” says CDSS interim executive director Laura LaChance. “For people with Down syndrome, it has the potential for creating greater independence. From daily reminders to keeping in contact with loved ones and accessing directions, voice technology can help facilitate infinite access to tools and learning that could lead to enriched lives.” 

CDSS and FCB have teamed up before on several highly praised campaigns. “Down Syndrome Answers” saw members of the community responding to commonly Googled questions about their condition and lives. “Anything But Sorry” revealed the one word parents of kids with the syndrome never want to hear. Most recently, the team tried to get the entire community on the global Endangered Species List as a means of advocating for more resources.

Each of these efforts saw people with Down syndrome take charge of their own destinies and work toward a better future. This latest project empowers them to improve Google’s technology so they can access its features and functions and improve their daily lives, says FCB co-creative chief Nancy Crimi-Lamanna.

Below, she addresses some questions about “Project Understood” in a conversation that has been edited and condensed.

Muse: Can you talk about the genesis of the idea?

Nancy Crimi-Lamana: Our desire was to get back to doing work like “Down Syndrome Answers”—work that could tangibly help the community in a moment that’s unique to them. That’s when it struck us: Right now, everyone is talking about voice, and how voice-activated technology will be everywhere in the next few years. But for a group of people who struggle with atypical speech and speech impairment, that technology is unusable. They’re being left out of the voice revolution, and yet, they’re the ones who stand to benefit from it the most. For people with common speech, it’s just a convenience, but for a community that struggles to live independently, it can make a meaningful impact. That became our seemingly impossible brief: How could we get voice technology to understand people with Down syndrome?

From there, what was the process forward?

The next few months were a blur of research and calls with various tech companies and platforms. The breakthrough came when we reached out to Google. They were open to talking to us, as they wanted to solve the issue as well. Our mission was further championed by our global FCB chief innovation officer, Kris Hoet, who put us in touch with an advocate at Google who has non-standard speech herself.

How many Down syndrome people participated in the initial test?

The initial trial was made up of nine people living with Down syndrome. This was a very tense time for us, as it would tell us if our ambition was even possible. Speech recognition algorithms work on recognizing patterns. The trial would tell us whether there were enough patterns in the speech of people living with Down syndrome for it to be able to learn and adapt. It took a little over two weeks to get the answer we had all been hoping for. We’re still in the early research phase, but the more voice samples we have, the more likely Google will be able to eventually improve speech recognition for everyone.

How many voices will you need altogether?

We’re aiming to collect 1,000 voices to start, but how long or how many voices it will take is still unknown, as every voice has unique nuances.

Can you walk us through how the recording process works?

Once they’ve been signed up, Google provides each participant with a web link to ChitChat, a voice recording platform, where they’ll be asked to record an initial set of 29 phrases and words which represent a variety of vocal sounds and patterns. This sample will be analyzed to do any technical troubleshooting and determine if the individual is eligible to record the full phrase set, which includes 1,700 words and phrases. This doesn’t have to be done all at once, but can be saved and completed [at the participant’s convenience]. The total time it can take to record this full set is between four and seven hours depending on the individual.

Some examples of phrases are:

• “He said buttercup, buttercup, buttercup, buttercup all day.”
• “I owe you a yo-yo today.”
• “Buy Bobby a puppy.” 
• “Bamboo walls are getting to be very popular because they are strong, easy to use, and good-looking.” 
• “Strawberry jam is sweet.”
• “She looked in her mirror.”

Why is voice technology an especially important resource for the Down syndrome community?

It can help facilitate classroom learning, employment and even relationships. As we move to a voice-first world, it’s important that everyone is included in this.

CREDITS

Campaign: Project Understood

Client: Canadian Down Syndrome Society
Chair: Ed Casagrande
Interim Executive Director: Laura LaChance
Marketing & Communications Manager: Kristen Halpen
Board Member: Ben Tarr

Google
Technical Program Manager: Bob MacDonald
Technical Program Manager: Pan-Pan Jiang
Product Manager: Julie Cattiau
Engineer: Jimmy Tobin

Creative Agency: FCB Canada
Chief Creative Officer: Nancy Crimi-Lamanna & Jeff Hilts
President: Bryan Kane
Associate Creative Director: Elma Karabegovic
Associate Creative Director: Michael Morelli
Associate Creative Director: Marty Hoefkes
Copywriter: Shannon McCarroll
Copywriter: Jason Soy
VP, Managing Director: Tracy Little
Group Account Director: Blake Connolly
Account Supervisor: Olivia Selbie
Agency Producer/s: Sarah Michener/Kristine Lippett
VP of Operations: Shandi Horovitch
Project Manager: Cori Pettit
Chief Strategy Officer: Shelley Brown
Director of Strategy: Eryn LeMesurier
Director of Strategy: Shelagh Hartford
Strategy Coordinator: Audrey Zink
Director, Product and Technology Solutions: John Sime
EVP, Head of Global Innovation: Kris Hoet

PR: Shannon Stephaniuk, Glossy

Production Company: Radar
Director: Scott Drucker
Line Producer: Sarah Michener
Director of Photography: Scott Drucker & Chet Tilokani
Camera Operator: Scott Drucker & Chet Tilokani
Audio: Nicolas Field
Hair & Make-Up: Neil Silverman
Photographer: Cassidy Clemmer

Editing House: Outsider Editorial
Editor: John Gallagher/Michael Barker
Editorial Assistant:  Scott Edwards
Executive Producer: Kristina Anzlinger
Transfer Facility: Alter Ego
Colourist: Eric Whipp
Online Facility: Alter Ego
VFX Artist: Eric Perrella
Alter Ego Producer: Caitlin Schooley-Groneveldt
Music House: Grayson Matthews
Music Track Director: Mark Dominic
Engineer: Vlad Nikolic
Audio Producer:  Kelly McCluskey

Speech Pathologist: Amanda Cotton

Website design: Bliss Interactive
Kris Van Wallendael
Julie Post
Dao Tran