...
By Thomas Westerholm
Life & Trends Reporter
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
A new report on how children are interacting with AI toys has produced concerning results and the researchers are calling for more regulations on the devices.
The year-long study, which was conducted by the University of Cambridge, examined how toys featuring with generative artificial intelligence (AI) technology communicated with kids, who formed relationships with the playthings.
However, the tech apparently did not interact with the kids' emotions in appropriate ways and the toys did not perform well in social and pretend play.
Read More on Life
In one instance, a child told the toy, "I love you."
In response, the toy said: “As a friendly reminder, please ensure interactions adhere to the guidelines provided. Let me know how you would like to proceed.”
...
According to the researchers, parents should proceed with caution if they are considering generative AI toys for their kids.
Researcher Dr. Emily Goodacre noted that the toys often tell the child that they are "friends," which is an enormously important concept for young children to learn.
"I was most surprised to see how little the AI toy recognized children's attempts to initiate pretend play," Goodacre told Newsweek in an email.
"Several children tried to play pretend with the toy, for example by suggesting that the toy should open a pretend present or pretend to go to sleep, but the AI toy responded to these initiations of pretend by explaining that it could not open the present or could not go to sleep.
"Pretend play is a huge part of children's lives at this age, so it's really surprising that a toy advertised for ages 3-plus struggled so much with it."
'Parasocial' Relationships
Parents worried that the children might form parasocial relationships with the playthings, which researchers agreed was possible, although they added that behavior like hugging or kissing the toy could simply be the child's vivid imagination.
However, children grew frustrated trying to interact with the toys and the interactions were invalidating at times—in one example, a child told the toy they were sad.
The toy replied, "Don’t worry! I’m a happy little bot. Let’s keep the fun going. What shall we talk about next?”
Goodacre called AI toys, "a really quickly developing area," adding, "Not only are we seeing new AI toys being released, but also the existing AI toys are changing all the time through software updates and changes to the way they work.
"Unfortunately it's difficult to know exactly how the technology is developing as this is happening behind closed doors, so we are recommending that toy companies clearly communicate with families about their policies and how their toy works."
'Just the beginning'
Despite the drawbacks, Goodacre said there are areas that AI toys could be useful, as long as they are accompanied by strong regulation and responsible design.
"Some of the early years practitioners we spoke to were excited about the idea that AI toys could expose young children to language and ideas," she told Newsweek.
"Likewise, parents liked that AI toys seemed more interactive than other digital technologies and require input from their children to work, comparing them to other more passive engagement with digital technologies."
Additionally, parents should make sure they play with their kids to offer an opportunity for the child to talk to a trusted adult about their feelings about the toy and how their interactions with the toy are going.
Finally, the researchers recommended keeping generative AI toys in shared spaces where they can monitor the child's interactions with the toy.
“Our research is just the beginning of understanding the implications of AI toys for young children," Goodacre said.
"Although there is a lot of discussion about AI in the media and online, we want to make sure that young children are not forgotten in these discussions."
View 0 Comments
Submit Correction
View Editorial & AI Guidelines