The expert. Jean-Gabriel Ganascia, an expert in artificial intelligence and machine learning, is a professor of computer science at Pierre and Marie Curie University in Paris, head of the ACASA (Cognitive Agents and Automated Symbolic Learning) team at LIP6 (Paris VI University’s computer science laboratory) and deputy director of Labex OBVIL observatory. His research now focuses on the literary aspect of digital humanities, computational philosophy, and the ethics of information and communication technologies..
What role could artificial intelligence play in children’s education and development through “cognitive” toys?
Jean-Gabriel Ganascia: artificial intelligence has already been incorporated into toys for a while, but on a small scale. Take dolls that respond to commands or interactive books that develop the story around the reader’s answers. This is a natural trend. Artificial intelligence creates a kind of much more flexible encyclopedia. The main difficulty with children is that they can ask all sorts of questions and sometimes in a very surprising way. The real challenge for artificial intelligence is how to answer while reformulating questions to adapt the register to the child’s age and level, then providing information that the child can take in. One thing is certain: a machine is always going to be more patient than a human being! I think that cognitive toys poucould find a use among some shy children who would find it easier to trust a machine. More generally, artificial intelligence could support children in their extra-curricular learning by adjusting the difficulty of the exercises in a more tailored way than they get in class and more effectively than their parents – ¬who can sometimes be out of sync with educational requirements – could do. A machine can be more attentive to errors and more specialized in its educational approach.
Further reading: Will intelligent objects take on a life of their own?
In 2011, IBM’s supercomputer Watson proved itself by beating two human rivals on the live TV quiz show Jeopardy. © IBM
Critics of cognitive toys highlight the dangers linked to exploitation of the personal data collected. Are they right to do so?
These toys’ design has to adhere to regulations on confidentiality and data protection. They will have to obtain certifications to guarantee this and move onto what we call privacy by design. Manufacturers have to incorporate this if they don’t want to give themselves a bad image. And lawmakers will doubtless have a role to play. To widen the debate, I would say that our societies are confronted by three conflicting aspirations. The first is a fierce desire for intimacy and respect for our private lives. The second is a demand for safety and protection against disease and terrorist risks (which requires us to hand over data). The third is an insatiable need for full transparency, which means we want to know everything about everyone. All these aspirations clash with each other and we will have to find a balance.
Dino Cognitoys Having demonstrated the performance of its Watson supercomputer, which managed to beat humans on the TV quiz show Jeopardy in 2011, IBM is working to find commercial outlets for it. To this end, the US company launched a competition to encourage developers to create apps that use Watson’s capabilities. This is how the Dino Cognitoys project came about. These little plastic dinosaurs can have conversations, joke and sing with children, getting to know them during the process. The toy is fitted with a voice recognition system and permanently connected to a cloud computing app based on Watson. The idea is that it becomes a companion that can constantly develop and “grow” alongside the user.
Will artificial intelligence become a consumer product?
Well, it already is a consumer product. Think of online search engines and voice assistants in smartphones. Artificial intelligence already plays a very important role in our information society, because the entire current business system is based on profiling tools. Artificial intelligence underpins the analysis of the masses of data every individual produces ? what we call big data. Product suggestions based on what we are more likely to want and recommendations are commercial tools based on artificial intelligence.
North American brand Mattel aroused debate after the release of its new Hello Barbie doll fitted with a passive voice recognition system that constantly listens to surrounding noise and creates a profile according to children’s tastes. In Germany, the press went as far as to rename the toy “Barbie Stasi” in reference to the former East German regime’s political police. © Mattel
Aside from toys, which general public sectors will artificial intelligence have an influence on in the near future?
A better question might be to ask which sectors artificial intelligence won’t influence! I would say transport is a key sector, with robotic driverless, and so is health, where the digitalization of data is a real revolution. We have sensors that can collect data constantly and we will see the arrival of connected devices to help carry out diagnosis and treat certain illnesses. Machine translation systems, which are becoming ever more effective, also come to mind. Artificial intelligence is going to play a bigger role in our lives. Whilst remaining invisible, it will transform everyday things such as cars, household appliances and home automation systems.
On the same topic
Will intelligent objects take on a life of their own?
Bees’ sense of smell can detect serious diseases
25 petabytes: how EDF is approaching big data?
The self-driving car is revolutionizing the automobile… And our world
In partnership with