Renowned artificial intelligence (AI) specialist, Yann Le Cun is a professor at New York University and head of AI research at Facebook. In 2016, he held the “computer science and digital sciences” chair at the Collège de France. He notably wrote When the machine learns. The revolution of artificial neurons and deep learning (Odile Jacob, 2019).
How do you respond to the many criticisms that flourish about the potential effects, imaginary or real, of AI?
There are many positive effects that people ignore, some are potentially negative and others are fantasy. Today, when you buy a car, it incorporates a system called AEBS [Advanced Emergency Braking System], based on AI methods, which allows for example to detect that a pedestrian crosses the street in front of a car and which automatically causes its braking to avoid the accident.
AI increases road safety in this way. According to statistics, this reduces collisions by 40%. Similarly, the use of AI for the analysis of medical images allows the detection of tumors in mammograms. In the coming decade, we can envisage the development of many applications in science, medicine and biology.
Thanks to AI, companies like Facebook and Google can also automatically filter out abusive content – pornographic, violent, hateful, terrorist in nature, etc. It's about allowing people to communicate without killing freedom of expression, while maintaining a certain civility in speech. There are a lot of positive applications in information retrieval, translation, natural language comprehension, speech recognition, Alexa-like intelligent virtual agents.
Where do you stand between those who fear a Terminator-style evolution and those who predict that AI will solve all problems?
Beneficial applications will far outweigh it. If some have negative effects, it will only be temporary. Any new technology creates perverse effects which are then corrected.
Currently, two concerns are perceptible in Europe. The first relates to employment and concerns employees who wonder how a robot can replace them in their work or make their professional skills obsolete. The second fear is the opposite: if Europe does nothing, China or the United States will take control of the economy, because these countries are more ahead of us in AI. Faced with these questions, the answer is both political and economic.