Voice (What Is It Good For?)
Since the debut of the Amazon Echo in 2014, smart speakers have revolutionized the way consumers use voice applications in their daily lives. As of early last year, nearly 90 million people in the United States alone use a smart speaker. Voice assistants are now found in our phones and homes, and are entering other channels, such as cars and stores. Despite this explosion of use cases for voice first technology, its potential has not been fully realized within certain populations.
While nearly all of us can benefit from a virtual assistant in some way, the best use cases are still being developed. As conversational AI advances, VUI (voice user interface) applications become particularly valuable for three demographics: children, the elderly and people with disabilities. Each of these groups presents opportunities to use voice applications to:
- Learn how to converse with people (and technology)
- Complete tasks more efficiently than a GUI (graphical user interface)
- Engage their attention through personalized feedback
Children Are the Future
Think back to your favorite toy as a child. Maybe you would narrate its actions with a particular voice or it came with a button that gave it a voice of its own when pressed. Fast forward to today and this capability is being developed on a much larger scale. Companies like Creativity Inc. are integrating voice technology in a suite of products from stuffed animals to toothbrushes to musical instruments. There are also voice apps that guide kids through interactive “choose your own adventure” storytelling games.
Beyond the fun that voice first technology brings, it is also a valuable educational tool. One example of this is through dictation, or speech-to-text, technology. While many devices have this functionality built-in, companies like SoapBox Labs have taken this a step further by developing a speech recognition API delivering immersive learning experiences that provide literacy assessments, dyslexia screenings and more.
Another compelling use case is how voice first tech teaches language. Bamboo Language illustrates the power of voice-based learning through its wide range of applications for K-5 students. As the pandemic continues to create inequality in the education system, it is reassuring that companies like these are finding creative ways to supplement time lost in the classroom.
Design for Your Grandparents
With a growing elderly population, it is becoming essential to design for more than the tech-savvy younger generations. China’s tech industry is going for the “silver hair economy” by creating products that are tailored to older people’s needs. This includes tools with simplified interfaces, for example, easy to use photo editing apps that allow sharing on social media without a caption.
“If the tech companies can design more products that meet old folks’ needs, with enlarged text, streamlined functions and simple design, that’d be wonderful.” — Fang Huiling a “techy” Chinese senior
One difficulty that elderly people run into (as do many people) is having to use a screen. Text may be too small, navigation too cumbersome, or your eyes just need a break from the pixelated glow. What if we could design products without screens? Enter products like ElliQ, which designs companion robots built for seniors. By suggesting personalized activities and engaging in conversation, these robots keep their users cognitively sharp. The market for robotic caregivers is ripe with possibilities, but even low tech Alexa Skills offer many of the same benefits.
The power of AI is also being tested in end-of-life care settings. These systems use algorithms to inform physicians when to have serious conversations with critically ill patients. Such a delicate use case begs the question of how reliable, or rather how we should interpret, data driven suggestions from our AI products. However you frame it, today’s AI is helping to facilitate important conversations between people across generations.
Accessibility is Better for Everyone
As a core design principle, accessibility should not be seen as a barrier to innovation, but rather a requirement for it. Are you bedridden but want to manage your regular household items? Smart home technology relies on interconnected products that typically use voice as the remote control. Other voice first products are going beyond helping people with vision impairments and limited mobility. Smart speakers are also teaching people how to talk, highlighting a significant use case for those with cognitive disabilities.
In order to create these advanced devices, voice assistants need to be trained on large datasets. This means if a product is trying to learn speech impediments, it must be trained to recognize these sounds and their context. However, there are studies that show automated speech recognition (ASR) systems are not being trained adequately. In particular, there are large racial disparities in the performance of these systems in recognizing speech. These systems can also have difficulty with people who are not native English speakers.
The lack of accuracy in ASRs demonstrates that AI is being trained with data that are not diverse or representative of its users. There are many discussions today about the ethics of AI and how we must be conscious of its design for use cases. It is a matter of time for technology to become more conversational and more human in its use cases.
As AI develops, it is more important than ever that designers focus on creating accessible products, with an emphasis not just on what the products will do but who will use them.