You may have already lost your balance at the top of an icy slope, knowing that your slide could end very badly.
For some, due to rapid developments in artificial intelligence, humanity currently finds itself in this very uncomfortable position.
This is feared by entrepreneur Elon Musk, Montreal (artificial intelligence) researcher Yoshua Bengio and many other experts who have signed a letter on this subject which is currently fueling the debate – with good reason.
They believe that large companies that design systems “more powerful than GPT-4” (the latest creation from the firm Open AI) should immediately put their work on hold for at least six months.
And if these companies refuse to stop training these systems, governments should impose a moratorium on the research, they argue.
This is not the first time that internationally renowned experts have sounded the alarm about the potential risks of artificial intelligence.
Five years ago, Elon Musk declared that “the danger of artificial intelligence is greater than the danger of nuclear warheads”.
At the time, his remarks raised some debate. But one could still think that these were whimsical statements from an eccentric entrepreneur.
This time, we are forced to take this kind of remark seriously. The speed of progress made in AI in recent months portends rapid changes for our societies.
Some are going to be positive, that’s obvious. But others, as we can already see, do not bode well.
Over the past few days, an almost surreal photo of Pope Francis has gone viral. The head of the Catholic Church is dressed in a pristine white down jacket.
PHOTOMONTAGE PABLO XAVIER, CREATED WITH MIDJOURNEY
Photo illustrating Pope Francis wearing a white coat created with Midjourney software
He looks like a Michelin man.
But the pope has never worn such a coat. The photo is fake – even though it fooled a lot of people. It is a man from Chicago who designed it with the Midjourney artificial intelligence software.
The previous week, it was photos of Donald Trump’s arrest that had caused a stir. False, too.
These examples are the tip of the iceberg on which we are embedding ourselves.
On this subject, the recent report by our journalist Charles-Éric Blais-Poulin was particularly disturbing. “Common notoriety individuals” in Quebec have been wrongly identified by ChatGPT as having been the subject of sexual misconduct allegations, he – notably – revealed.
These technologies (some generating images, others text) are now available to everyone, which means that we are entering a new era of misinformation.
Trust, the cement that holds our societies and their democratic institutions together, has already been shaken. But we may not have seen anything yet.
And we’re not talking about all the other potential risks associated with AI, ranging from discrimination to lethal autonomous weapons, to the survival of our jobs.
The fundamental problem: there is not yet a sheriff to oversee the development of artificial intelligence.
At least in Canada the process is underway. A bill (C-27) is at second reading stage. Yoshua Bengio believes that things are going “in the right direction”.
But until the legislation is passed, here like everywhere else, it’s the Wild West.
It is therefore urgent to regulate artificial intelligence. But a break?
“Of course I think it’s great to take a step back, but many people have been saying it for a long time and have sounded the alarm. Work has shown the societal impacts of artificial intelligence models. But it’s as if it had gone under the radar, ”explained the director general of the International Observatory on the societal impacts of AI and digital technology, Lyse Langlois.
At that time, there had been no break.
There won’t be any more today.
No offense to the signatories of the letter, it is a wishful thinking.
The competition between large American corporations, but also between the United States and China in this area, is too strong. The stakes are too high.
However, that does not mean that the initiative of Yoshua Bengio and his peers is a shot in the dark.
On the contrary.
They play the role of whistleblowers.
They have the merit of having made us reflect even more on the importance of marking out as quickly as possible the “uncontrolled race” in AI which we are still witnessing, alas, as too passive spectators.