Hey Siri, want to be friends? Children talk to tech, but does it respond?

A University of Washington (UW) study has examined what authors describe as “a missed opportunity” for voice-activated devices, or digital assistants, such as Siri and Alexa, to reach every member of the family.

Children communicate with technology differently than adults do, and a more responsive device – one that repeats or prompts the user, for example – could be more useful to more people, guiding children as they become more developed in their speech.

Co-author of the study Alexis Hiniker said “There has to be more than ‘I’m sorry, I didn’t quite get that’ – voice interfaces are designed in such a cut and dried way, and there needs more nuance.”

“Adults don’t talk to children and assume perfect communication,” Ms Hiniker said, “that’s relevant here”

With nearly 40 million US homes having a voice-activated assistant, such as Amazon Echo or Google Home, and with some estimates saying this will jump to more than half of US homes by 2022, the devices, and their role in helping children communicate and be understood, cannot be underestimated, authors said.

While some interfaces are specifically targeted to younger users, research has shown that the devices typically rely on the clear, precise English of adult users – and specific accents within that group.

Adult users who have English as an additional language, or even those native English speakers with a complex regional accent tend to “hit snags” according to a recent analysis by the Washington Post.

The UW study shows that children persist in the face of a communication breakdown, treating a digital assistant as a conversation partner, and, in effect, showing developers how to design technologies that are responsive to families.

Ms Hiniker said these devices are billed as “whole home” assistants, providing a central, collaborative and shared experience, however they all too often exclude children from this space.

In conducting the research, the UW team recorded 14 children aged three to five years of age as they played a Sesame Workshop game – Cookie Monster’s Challenge. As designed, the game features a cartoon duck who waddles across the screen at random intervals. The child is asked to make a quacking sound when they see the duck on the screen, and the duck is supposed to quack back.

For the purpose of the study, the duck did not quack back.

That scenario was something of an accident, Ms Hiniker said. The team, with funding from Sesame Workshop, was originally evaluating how various tablet games affect children’s executive function skills. But when they configured the tablet to record the children’s responses, researchers later learned their data-collection tool shut off the device’s ability to “hear” the child.

What the team had, instead, was more than 100 recordings of children trying to get the duck to quack to them – in other words, to remedy a lapse in conversation – and their parent’s efforts to help. From this, a study in how children communicate with non-responsive technology was born.

Researchers grouped children’s communication strategies into three categories: repetition, increased volume and variation. Repetition – in this case, continuing to say “quack,” repeatedly or after pausing – was the most common approach, used 79 per cent of the time. Less common among participants was speaking loudly – shouting “quack!” at the duck, for instance – and varying their response, through their pitch, tone or use of the word. (Like trying an extended “quaaaaaack!” to no avail.)

In all, children persisted in trying, without any evidence of frustration, to get the game to work more than 75 per cent of the time. Frustration surfaced in fewer than one fourth of the recordings. Surprisingly, in only six recordings, children went to an adult to ask for help.

When asked to assist, parents were happy to do so, but researchers found adults were also quick to determine something was wrong, and to encourage children to take a break from the game. Adults usually suggested the child try again, before trying themselves, and then pronouncing the game broken – only then did the children stop trying.

The results represented a series of real-life strategies families used when faced with a “broken” or uncommunicative device, Ms Hiniker said, whilst also providing a window into young children’s early communication processes.

As conversation partners with a child who is developing their speech, adults are good at recognising what the child wants to say, and filling in for the child, Ms Hiniker said, adding “a device could also be designed to engage in partial understanding, to help the child go one step further”.  

For example, a child might ask a digital assistant to play “Wheels on the Bus”, but if the device does not pick up the full title, it could respond with “play what?” or fill in part of the title, rather than saying “I’m sorry, I don’t see that in your playlist.”

Such responses could also be useful to adults, Ms Hiniker pointed out, saying “Person-to-person conversation, at any age, is filled with little mistakes, and finding ways to repair such disfluencies should be the future of voice interfaces.”

“Artificial Intelligence is getting more sophisticated all the time, so it’s about how to design these technologies in the first place,” she said. “Instead of focusing on how to get the response completely right, how could we take a step toward a shared understanding?”

An additional study has been launched by UW into how diverse, intergenerational families use smart speakers, and what communication needs emerge.

The study has been published in the proceedings of the 17th Interaction Design and Children Conference, held in June 2018 in Trondheim, Norway.