The Pretty Face of the Internet of Things
We have been looking at the implications of the interconnection of multitudes of devices – for security, for privacy, for property. What happens when the things connected with you and each other on the Internet can recognize your voice, and talk back to you? Voice recognition technology has made rapid progress, and it is already becoming normal that one can ask questions out loud to a computer (generally a “device”, a phone) and have it answer.
Thus Siri, Google Now and Cortana. Not only do they have access to a universe of information, but they learn about the users and refine their responses to suit the users’ needs and tastes.
In the movie “Her” (2013), set not too far in the future, the lead character falls in love with the personality of his personal device, and starts to believe that his feeling might be reciprocated, in part through the device’s (“her”) curiosity about him.
Now suppose this interactive voice-based capacity is put in reach of a child. What might a child believe is going on? What is normal to someone who is six or eight?
We don’t have to speculate. Mattel and ToyTalk now offer us – or at least the US – Hello Barbie, a talking Barbie doll. It will recognize its owner’s voice and communicate what the owner says to its large database. The database has a lot of information that it can transmit in child-friendly ways. It also learns the interests and tastes and personal information of the owner, so it can put its responses in a personal context for each child.
The doll can answer questions built on data gleaned from previous conversations with the child, such as “what should I be when I grow up?”.
In one demonstration it suggested being a dancer or a politician, or a “dancing politician” after learning that the child likes to be on-stage. (The Guardian)
No more need for an imaginary friend, when your doll can talk with you all you want.
Some people think that’s marvellous: your child won’t be lonely any more, or bored. Some people think that’s creepy: what does your child believe about her doll, and what is the corporation going to do with the information? Besides, isn’t it better for the development of creative imagination to have to overcome boredom with one’s own resources? (That line of inquiry could lead to interesting observations about adults too.)
How secure is the talking Barbie? The company claims that everything is encrypted from end to end. Further, the doll transmits (to a recorder) only when the belt buckle is pressed. Others have claimed that the doll can be hacked, and once it is hacked, the security and privacy measures are readily circumvented. It becomes a handy spying device in the hands of your kids.
The privacy concerns about what a child might say to a ‘best friend’ doll – all of which is sent to and recorded by the company – are serious. The company – rather the companies, Mattel the dollmaker and ToyTalk, the maker of the voice-activation feature and the keeper of the records – claim that no information derived from a child will be used to market anything to the child. The information is used only to enhance knowledge of the server, to improve voice-recognition techniques and to tailor requests to the individual.
However, the doll will ask questions about the girl’s favourite movies or fashions, and talk about such matters – and other Barbie products. Some people think that is a form of marketing. There are about 8000 different “conversations” in the Hello Barbie repertory, and those will grow as the girls confide in their dolls – it’s all grist for the “trend bucket”.
The data provides the creators with ideas for how to improve the product for the spring season, such as which lines they should add to Barbie’s response repository. So if the company notices a surge of kids mentioning Taylor Swift, Barbie may have some thoughts on the singer-songwriter a few weeks later. (Newsweek)
There is also talk of sharing information with third-party vendors for research and development.
Further, they will share the information with the child’s parent(s), on request. The parents need give their consent to the use of the doll – the US Children’s Online Privacy Protection Act (COPPA) is stringent about consent under the age of 13. The parents have to download an app and provide an Internet connection.
In addition, parents can review what the child is telling her doll. They can require the company to delete any recordings they wish to have gone. But is the sharing appropriate? May a child not have dreams or fantasies or conversations with a friend without the parents listening in? Will the child feel betrayed when she finds out that the parents have been listening? In short, the privacy protections come with points of caution or concern.
And suppose that the child talks to Barbie about being abused. Do Mattel and ToyTalk “know” what is being said? Are they subject to laws that require people who learn of abuse to report it to the authorities? Is it safe to pass on the recordings to the parents who may be the abusers?
Hello Barbie has provoked a fair bit of concern in the US, such as the lobby group “Hell No Barbie”. And since it’s in the US, there is a class action going on. No matter about the (parental) consent to the recording, the (parents of the) little friends of the doll owner have not consented, but they will predictably be talking to the doll too, or their voices picked up as the owner talks.
For the moment Hello Barbie is not available in Canada. Is this because of Canada’s privacy laws? Do we have anything stronger than COPPA? Would PIPEDA or provincial equivalents (where applicable) give kids a privacy right against their parents for listening into their conversations? It’s not really a commercial use for the parents to hear recordings of their kids, though Mattel’s distribution of the recordings to the parents is part of a commercial enterprise.
Or is it just staged marketing, and the talking Barbie will soon be on a toyshelf, or in a stroller, near you?
If so, not everyone will find her a welcome voice in the conversation. A number will find her to be one more example, perhaps, of what Professor Tom Keenan calls “Technocreep” – the apparently inexorable advance of creepy technology. Will the law stand on guard for us against the creepy? How?
[A tip of the hat to Jon Garon, Dean, Nova Southeastern University College of Law, for the title.]
And if this weren’t upsetting enough, there are new products that are even more insecure and whose uses of what is overheard is even broader.