“AI capable of disconnecting us? This is science fiction, “said Antoine Bordes

“AI capable of disconnecting us?  This is science fiction, "said Antoine Bordes

They are everywhere. On our computers. On our smartphones. A simple web search. An Automatic Translation Artificial Intelligence (AI) Are becoming more and more important in our lives. Without, necessarily, we are fully aware of it. They are mobilizing more and more experts around the world.

Is Stems, Brand new living room, World Artificial Intelligence Cannes Festival, By Larousse’s definition, “the ability to simulate human intelligence” is entirely devoted to these technologies. Organized this Saturday, at the Center for Congress in Crochet, the event is also open to the public.

Opportunity to calculate the growth of these AIs. How far can they go? What are their limitations? 20 minutes Asked researcher Antoine Bordes. Until 2014, at CNRS, this expert was the Associate Director of Fair (Facebook Artificial Intelligence Research), the basic research laboratory for artificial intelligence in Meta.

What does the birth of this new event in Cannes signify?

There were already programs related to the Mobile World Conference in Barcelona, ​​smartphones, CES in Las Vegas, or Vivatec in Paris. But the event, which took place in Cannes, was the first event dedicated entirely to artificial intelligence, from research to consumerism. The novelty of showing where it plays in the economy today.

How fast has technology grown in recent years?

There are still more complex parts for machines. But for some, semantic recognition, machine translation, and speech recognition are considered classics, and progress is actually quite fast. Much faster than experts predicted.

What are the applications used in AI today?

Let’s start from the opposite end and it will go faster. Today, of all the digital related, what are the applications that do not use AI? Very very low. It is used to create new experiences and new hobbies by providing new tools for creators and brands. It is explicitly used to organize sites like ours; For example, find content that is not there and protect them. The entire ecosystem is affected by AI to varying degrees.

AI also promotes science

It really helps a lot in scientific discovery. In biology, in chemistry. It enables very rapid advances in clinical research, data processing, or hypotheses.

I can give two examples. One of them is called Fast MRI. This is an algorithm that allows image reconstruction to speed up the medical imaging process, which can be lengthy and somewhat traumatic. Usually, this lasts 15 to 30 minutes. We divide this time into about four by taking fewer pictures and asking AI to complete the sequence. It was clinically examined by doctors. Like all the research we do it is completely open source and can be used by manufacturers.

Our colleagues at DeepMind [Google] They have also developed software that predicts the structure of proteins from their amino acid sequence. This will be the basic building block for tomorrow’s biology.

Let’s talk about the future. AI will continue to grow over time, but will there be a limit to its growth?

Yann LeCun, a researcher who works with us, often explains that AI is more like animal intelligence than human intelligence. That is, it may be more advanced in specific categories, but not in general. Doing certain types of tasks will be more and more efficient. On the other hand, it seems very complicated to me to see the same AI making a recipe, analyzing proteins or translating all languages. To get a brilliance that can be called “public” that can do everything, we are still far from it. So, yes, in that sense, I think there is a limit.

So is there no risk of seeing the machine gaining conscience and turning against its creator in some science fiction film?

There was a discussion on Twitter that there might be emotional AIs. But I think the community is still very skeptical about this. At the same time, we don’t really define what the machines will feel like … it has to be done. In the meantime, we are developing a regulatory framework to define their use, the data on which they are based, and the applications to which we have the right to use them. And an AI with an intelligence capability can go in a subtle way beyond its limits and disconnect us for sticking together … this is precisely science fiction.

Compared to this animal intelligence we are talking about, some people think that AI might want to defend itself without being disconnected in the name of survival instinct … which seems very imaginary to me.

Leave a Reply

Your email address will not be published. Required fields are marked *