Maintenance. "cinema helps anticipate artificial intelligence"

Cinema makes it possible to make more concrete the strange notion of artificial intelligence. By constantly imagining new possibilities that may not turn out to be so wacky in the future. Interview with Alexandre Pachulski, Doctor of Computer Science, author, lecturer, blogger and co-founder of Talentsoft, software publisher.

How can we simply summarize artificial intelligence (AI)?

It is to make machines do what requires human intelligence. A much wider field than computer science. It appeared in the 1950s, building on the cognitive and human sciences.

Doesn't cinema stop offering us different forms of AI, even before they exist?

Cinema, and more broadly science fiction (SF), has an ability to anticipate. With the endless question of whether the SF inspires the real or whether the possibilities of the real inspire the SF. But with 2001, Stanley Kubrick's Space Odyssey, at the end of the 1960s, we had already been able to imagine a so-called intelligent machine capable of piloting a mission, directing a spaceship and converting with astronauts. But the SF, 98%, approaches AI through dystopia (dark utopian world) and not utopias. Positively described AI is very rare.

Because to create a playwright, it's better if a train arrives late …

Writer and director Aaron Sorkin says that a good film is the tension between the search for the hero and all the obstacles facing him.

So cinema gives a generally negative idea of AI?

In Kubrick's 2001, we see a machine that considers humans to be obstacles to his mission. That's exactly what all our fears of AI are based on. Could a highly developed artificial intelligence consider that the human being is a hindrance to the development of humanity? When you look at the state of the planet, she might have that idea! In fact, when cinema incites us to ask ourselves societal, political, philosophical questions, it can inspire reality and enable us to make an AI that will serve us positively.

In Blade runner, filmmaker Ridley Scott wonders if robots can have a consciousness?

Researcher Yann Le Cun says that the best AI has the common sense of a 2-year-old child. At the same time, I studied artificial intelligence in the '90s and all my teachers swore to me that an AI could never win the go game. She did. If an AI can become aware of signals corresponding to our five senses, it can forge a form of consciousness. But it'll only be mechanical.

Unless you have AI with a biological share?

Even if one day she has a share of biology, there will always be a human behind who will program it, including for her to learn for herself. It is the human who sets the objectives and sets control mechanisms.

Many films and series talk about the use of AI in the service of security?

Cinema gives possible scenarios. In person of interest, an AI monitors every street in the world with the intention of protecting the poorest. A very nice intention. But behind them are the questions of data processing, possible deviations, the right of people to remove their data … In the real humans series, a nurse android refuses to donate coffee or chocolate in the name of health. In fact, AI must correspond to a social project built with citizens.

But political power can change …

That is why the only way out is for citizens to have the data collected by third parties. Otherwise, there will be endless blockages. Look at the stop covid app, we're still very bad because we Don't trust our leaders. It's a real political subject. The AI is an assistant who must be able to tell you why he gives you this opinion, as a friend does.

Social media algorithms, are they AI?

The goal of social media is to keep you at home. That's why, as soon as you've seen one video, they offer you another one of the same style. The risk is to lock you up in your own tastes. Finally, you are offered what strengthens you in your beliefs. The risk is that humans end up looking at the world only through what they already know. It points to the responsibility of these companies to push for a more ethical AI. There is no way out of technological issues without addressing education and social issues.

What if everyone could choose their algorithm, their AI?

We need to be able to think globally with respect for the individual. If we Don't think of the public interest, we run the risk of building highly individualistic societies.

What is the closest future to AI?

What is going to become more and more important is everything about consumption. And that's not good news. Otherwise, the AI will enter the labour and training market in an interesting way by helping to develop skills for new trades that do not yet exist.? And in the area of health, there will be a lot of things.

What's your favorite movie?

The first Blade runner. It's an aesthetic culmination with the music of Vangelis and extraordinary actors. With many exciting questions: what matters to love someone? And what would happen if we knew our expiry date?

I.A. Generation 80 films and series to decrypt artificial intelligence, Ed. E / P / A. 216 pages. 35 €.