– by Gry Hasselbalch
Toy manufacturers are today creating intelligent toys that remember, find patterns and respond to data from children. We need a data ethical approach to innovation in the development of an “Internet of Things” for children.
Remember E.T. ? A creature from outer space, but also the most amazing toy on earth, a real life friend to the boy Elliot, who finds him. In the 1980s Steven Spielberg created a career out of films based on children’s greatest fantasies. E.T. is a depiction of the ultimate toy that we have all dreamed about at some point in our childhood: a companion that learns with you, educates you, grows and develops with you and knows you well.
Today technology innovation is catching up with our dreams. The Internet of Toys operated by machine learning technologies and cognitive computing consists of toys that can actually act like children’s very own personal playmates.
The Cognitoy dinosaur is based on one of the most powerful machine learning technologies in the world, the Jeopardy winning IBM Watson computer. It remembers and learns from a child’s responses and tailors its own responses to these. The Fischer Price Smart Toy is an intelligent teddy that on the promotional website as described: “As unique as your child…It actually responds to what your child says and remembers things. It takes cues from him or her, then invites play, talk, movement, imagination and learning”.
A toy that remembers everything
The public response to these internet connected smart toys have been mixed. The Mattel Hello Barbie that recognizes and processes a child’s voice in a commercial cloud was immediately upon its release dubbed the “Surveillance Barbie” in the media, and the Google patent on a Teddy equipped with cameras in the eyes that follows the movements of its owner was seen as more “creepy” than cute.
Though in many instances we are wooed by the smart toys’ humanlike appearance and behaviour. They talk funny, look adorable, know knock knock jokes and your child’s favourite colour. They have eyes like E.T. with depth and sincerity.
But we need to remember that these toys are in fact more alien in nature than they are human. Their nerve systems consists of programmed algorithms. Data is the blood that flows through their veins.
They are built on a different type of intelligence, an A.I. created in our image, but still different from ours. A quantitative kind that analyses, finds patterns, acts on profiles built out of all the bits of data that children feed them with – the sound of their voice, their face expression, the content of their dreams and their wonders.
They learn and develop by remembering everything and so our E.T. is always with us, even when we grow up and place the physical toy in a box in the basement. It is there somewhere in the clouds, remembering you as a child.
They don’t belong to you either. You buy the toy, but you have no control over or insight into the interests that have guided the development of their nerve systems, and the data that you feed them with does most often not belong to you.
They also want to phone home to the proprietary cloud based platforms where data is stored and processed, so the toy can learn with the child, understand the child, respond to and guide the child.
Ask Questions to the Internet of Toys
A.I. is today a key area of commercial investments and innovation among the largest tech industry players. IBM, Google, Facebook, Apple (and the list goes on) all have larger sections dedicated to the development of machine learning technologies that can create meaning out of data, they extensively hire new staff in the field and invest great amounts in the it. Intelligent toys is a category of toys closely connected with a datadriven commercial sphere. We need to ensure an ethical approach to data innovation in the development of IoT for children. These are some of the questions we could ask (but we should ask many more):
- Data profiles: A smart toy is smart because it has a detailed data profile of a child. How can we ensure that children’s data profiles act on their behalf now and in the future? Where is the profile stored, who has access to it, how is it combined with other data?
- Algorithms: Algorithms act on, inform and direct the smart toy’s interaction with a child. How do we ensure transparency and ethical standards in the programming of algorithms for toys?
- The proprietary cloud: Implications of centralization and monopolization of data is as important as security. Who owns children’s data? How do we ensure that children have a “clean slate” when they grow up? That they and their parents have control over their data?
- Commercial interests: The smart toy is a personal playmate that educates and guides the child. We should demand transparency and ethical standards of the commercial interests embedded in e.g. the toys’ responses to the child. How do we ensure that the toy does not become an invisible marketing tool?
- Social implications: Parents most often will have access to the child’s private interactions with the internet connected toy. How do we preserve a child’s right to privacy in the social sphere?
- The meta narrative: There is always a risk of manipulation and control embedded in data driven technologies. What type of meta narrative do we want to drive innovation in technologies for children? Do we really want detailed digital data profiles on children?