AI “is a nice crutch,” says lecturer Samuel Nowakowski, “but we’re going to have to constantly check that it’s going to support us or at least perform the task for which it was designed.” Photo: LSF2024/Youtube

AI “is a nice crutch,” says lecturer Samuel Nowakowski, “but we’re going to have to constantly check that it’s going to support us or at least perform the task for which it was designed.” Photo: LSF2024/Youtube

How can we occupy the terrain of artificial intelligence? Samuel Nowakowski shared his perspective on Tuesday on the stage of the Luxembourg Sustainability Forum, organised by IMS at the Tramschapp.

“If there’s one thing that machines will never have, it’s stage fright. Even though I’ve had a long career as a teacher, being in front of an audience is never something you take for granted…” Samuel Nowakowski--lecturer in information and communication sciences and researcher at the Lorraine Computer Science and Applications Research Laboratory (Laboratoire lorrain de recherche en informatique et ses applications) at the University of Lorraine, the Nancy Ecole des Mines and Sciences Po--may be used to ‘evangelising’ on his favourite subjects, but he is still a man, and that is what he reminded the audience at the Luxembourg Sustainability Forum 2024 on Tuesday 22 October at the Trampschapp.

“I teach digital humanities! What kind of oxymoron is that? It’s simply about getting people to think about the world we live in, and, in particular, the world in which the digital dimension has become so important that we are no longer the same human beings we were 30 years ago. A whole host of tools are chasing others away at an ever-increasing pace,” Nowakowski, who has worked with Steve Jobs, for Canal+ and for the French government, explained to an audience of experts and high school students. “We talk about artificial intelligence with these young students who will one day be decision-makers, and I give them a lesson on how it works, what the different elements are.”

And, in a confident tone, he continued: “One of these young people said to me: ‘I have experience with AI.’ He’s a scout and, at his scout camp, one day he was asked to prepare a mass. As he didn’t know how to do so, he asked ChatGPT to produce a mass. When he started to look at what was in the text, the system invented or rewrote passages from the Bible. Because of a lack of information, ChatGPT went looking for elements in the texts that were closest and most relevant from a probability point of view... The content was completely crazy!”

“Behind this little anecdote, there is an enormous amount of information. The system can answer all the questions, but only human criticism can detect what’s going wrong. This is true for my student, but also for lawyers or programmers... We are faced with systems that put us back in a position to question the production of knowledge... Then, it also questions the confidence we had in an automated machine to which we entrusted a task, as in the production of cars.”

“Today, you have something that will produce information, content, a text, and so on, with a degree of probability of confidence. You’re not necessarily going to know whether it’s 90% reliable, 80% reliable or only 50% reliable. It’s a nice crutch--which looks like a crutch--but we’re going to have to constantly check that it’s going to support us or at least perform the task for which it was designed.”

“Human element is extremely important”

“I’m not in the process of saying that everything should be thrown out,” said Nowakowski, who, like the rector of the University of Luxembourg, Jens Kreisel, is aligned with the idea of interdisciplinarity. “The question is rather what we do with all this. We need to put the human element back in, and the human element is extremely important in terms of discerning what is valid and what is not. It’s the ability to be critical.”

His ten-minute conference went by too quickly and ended with a more political message. "For me, it’s an act of citizenship! It’s both a question of profession--I have the skills and it’s not the machine that’s going to dispossess the human being and the one who has the last word. In medicine, in theatres of war--which are terrible, I agree--the human being has the last word. In democracy, there’s the assumption that citizens are capable of understanding. If I want to be able to debate AI, its place and the society we are building with it, I need to be able to understand what I have in my hands and retain the skills we have across the whole spectrum of human inventive activity. A machine will never be able to invent. Democracy isn’t just about slipping a ballot paper into a ballot box, it’s also--and above all--about being able to question and question with common sense the major issues facing society.”

This article was originally published in French.