Are robots about to rule the world 2023 ?
Are robots about to rule the world 2023 ?
Are you ready for the revolution? Recent advances in artificial intelligence, as exemplified by the successes of the ChatGPT chatbot, are forcing us to think about our future. Will we take change by the hand, or will it take us by the throat? This is what our editorial writer Alexandre Sirois wonders about. One thing is certain, admitting defeat would be the worst strategy…
This is no longer science fiction.
“As soon as artificial intelligence makes better decisions than us about our careers and relationships, we will have to rethink our conception of humanity and life. »
Israeli historian Yuval Noah Harari predicted this five years ago in a book called 21 Lessons for the 21st Century.
He plays the futurist in particular in the first chapters of this book, which deal with “the technological challenge”. He sees a future where humans rely on algorithms to decide which discipline to study or which company to work for.
And even to choose the people they will marry!
In short, he describes a world where “authority” would pass from humans to computers.
I found these chapters fascinating when I read this book a few years ago.
But I tended, at the time, to associate these ideas with science fiction.
They found themselves in the same category as the futuristic scenario of 2001: A Space Odyssey. In this feature film by Stanley Kubrick, a computer equipped with artificial intelligence takes control of a spaceship and tries to eliminate its crew members.
However, like many, I have tested the ChatGPT conversational robot over the past few weeks. And now there’s more science and less fiction than I thought there was a historian Yuval Noah Harari’s words.
Are we really about to witness a revolution? To what extent will the coming major changes shake our society? How will new AI-powered tools impact our decisions? And, more broadly, on our lives? I tried to find out by talking to a dozen experts.
Starting with Yoshua Bengio, a full professor at the University of Montreal and a world authority in artificial intelligence.
PHOTO ALAIN ROBERGE, LA PRESSE ARCHIVES
Yoshua Bengio, full professor at the University of Montreal and world authority in artificial intelligence
“Can we speak of a revolution, Mr Bengio?
– Possibly, yes. It will happen gradually. Artificial intelligence will transform not only our economies but also our daily lives. And even more profoundly, the idea we have of what a human being is, what intelligence is, and even consciousness. It is radically transformative. »
A pivotal moment
The computer scientist is not surprised to see how the appearance of ChatGPT was seen as a pivotal moment in the development of artificial intelligence.
He points out that science advances in small steps. And that, at a certain moment, one of these steps is considered fundamental. He compares the phenomenon to a vase filled with water, one millimetre at a time. One day it will overflow.
“At some point, it becomes usable. We reach a certain level of performance, meaning we will find that in our telephones. We pass a threshold that makes companies want to take it because they see that it is good enough to bring added value,” he describes.
Are robots about to rule the world 2023 ?
With ChatGPT, therefore, the vase begins to overflow. Moreover, even if this robot is still in development and its current version is imperfect, it enjoys phenomenal success.
It passed the milestone of one million users five days after its launch in November 2022. Some on social networks did not fail to recall that it took Facebook ten months to do the same, five months at Spotify and three and a half years at Netflix.
But what is ChatGPT for?
Let’s ask him!
“What is your use, ChatGPT?”
– I can be used for many tasks related to natural language processing, such as answering questions, translating, generating text, automatic writing, understanding voice commands and much more. I can help people access information quickly, automate some data entry tasks, and improve the quality of communication. Companies can also use my abilities to improve chatbots and virtual assistants and create more sophisticated recommendation systems. »
Are robots about to rule the world 2023 ?
Impressive. You will surely be amazed whether you test its creativity or efficiency. As much for a poem on climate change in the style of Gilles Vigneault as Spanish lessons or a summary of the October crisis!
Most chatbots can line up pre-programmed sentences. But ChatGPT does much, much better. It analyzes natural language and can recreate it to answer our questions, thanks to neural networks that work like the human brain.
Impressive for everyone
The ingredients of its success? We asked Joëlle Pineau, director of artificial intelligence research at the Fair Laboratory of Meta (formerly Facebook), to explain it.
She says, “We think it relies on a lot of data, which is amassed directly from the internet,” but also from other sources such as books and podcasts.
However, the human work behind the machine should be considered.
Joëlle Pineau, director of research in artificial intelligence at the Fair de Meta laboratory
These humans will annotate conversations to indicate good, less good, or even toxic.
Joëlle Pineau, director of research in artificial intelligence at the Fair de Meta laboratory
Humans also watch over the classification of the answers so that the system offers the best possible ones.
Olivier Blais, co-founder of Moov AI
“ChatGPT is one of the few artificial intelligence tools that are as impressive for the average person as it is for the expert. It’s something revolutionary,” says Olivier Blais, co-founder of the Montreal company Moov AI, which offers artificial intelligence consulting services.
He speaks with the enthusiasm of a child in a candy store.
We have already reached the point where artificial intelligence “will provide information to make informed decisions. And it will get even better,” he said.
As several experts consulted, he spontaneously cites the health sector when asked for concrete examples to understand who could benefit the most from this revolution.
Are robots about to rule the world 2023 ?
He imagines very well that a tool like ChatGPT could be responsible for the consultation service of the 811 lines.
“Right now, if you’re using ChatGPT, it’s very general and not necessarily reliable. On the other hand, when we adapt it to a corpus, it will be different. For health care, we could provide him with all the 811 transcripts” and thus obtain “more convincing and more reliable” results.
A robot to replace pros
For this reason, almost everyone who knows anything about artificial intelligence agrees that the world of employment will be turned upside down.
Even since the appearance of ChatGPT, the status of professions, we believe to be safe now seems more precarious.
For example, it has long been feared that smart cars could put taxi drivers out of work. Or that the artificial intelligence applied in the shops makes a good part of the humans disappear (on vacation in Seattle with the family last summer, I visited an Amazon Go grocery store and did not need to check out to pay).
But here, the list just got longer. Will we now need fewer doctors? Fewer journalists? Fewer teachers?
Will the machines, from the top of the knowledge they will have accumulated by being fed astronomical quantities of data, come to make better decisions than us?
“I assure you, we are not there yet,” replies Foutse Khomh, full professor in the computer engineering and software engineering department at Polytechnique Montréal.
Foutse Khomh, Full Professor in the Department of Computer Engineering and Software Engineering at Polytechnique Montréal
The risk of being replaced in the short term, at least for tasks that require our creativity or our understanding of the complexity of human and social interventions… we are still far from that.
Foutse Khomh, Full Professor in the Department of Computer Engineering and Software Engineering at Polytechnique Montréal
Are robots about to rule the world 2023 ?
But advances in this area are pushing Professor Khomh, like the director general of the International Observatory on the Societal Impacts of AI and Digital Technology, Lyse Langlois, to sound the alarm. During the virtual interview, they gave us, they both insisted: we must ensure that development is done in a “responsible” way in artificial intelligence.
But that is, of course, much easier said than done.
When artificial intelligence goes wrong
Assistant professor in the management department of Laval University, Gaëlle Cachat-Rosset warns us at the start of our interview: she will submit us to a test, as she would with her students.
She asks us to enter the words “a nurse” in Google translation and translate them into… Hungarian!
The result: nővér.
Then she tells us to try translating it again, this time back to French.
Strangely, Google’s machine translation tool, powered by artificial intelligence, offers us the word sister. We could also have obtained the word nurse, explains the professor.
Are robots about to rule the world 2023 ?
“It gives a feminine! You started with a masculine word that you translated into a language without gender pronouns. And when we retranslate from Hungarian, we see how gender stereotypes associated with the profession appear. »
Gaëlle Cachat-Rosset specializes in research on issues of diversity and inclusion in organizations. With the use of artificial intelligence becoming widespread, she has her work cut out.
Amplified biases
She explains that we have too little awareness of the risk to which we expose ourselves by letting artificial intelligence make more decisions for us.
And she tells us that there are already many examples of abuses related to discrimination, biases and stereotypes perpetuated by artificial intelligence-based systems like ChatGPT.
One of the most (sadly) famous is that of Amazon, which decided to recruit employees using artificial intelligence in 2014. However, the digital giant quickly found that this system excluded women in the process for positions like IT developers. He had been fed the CVs of previous employees, mostly men. He concluded that a good employee had to be a man.
Are robots about to rule the world 2023 ?
When artificial intelligence learns historical data from an organization or our society, it tends not only to reproduce the existing biases but also to amplify them because it seeks the straightest line to have a successful result.
Gaëlle Cachat-Rosset, assistant professor in the management department of Laval University
Another expert, Sandra Rodriguez, also gave us an example of poor recruitment at Amazon. This new media sociologist is also the creative director of immersive artificial intelligence works. She has just designed one to help us “learn what’s under the hood” of this new Technology.
How we make artificial intelligence, in short.
This project, which the National Film Board of Canada will present in Montreal in a few months, allows you to converse with a digital double of the famous American linguist Noam Chomsky.
A conversational robot was therefore created from “digital traces” of this expert found on the web, including numerous texts and public interventions. This robot also allows those questioning it to understand that all its answers “are led by algorithms”. And see how they are created.
Are robots about to rule the world 2023 ?
One thing leads to another, our conversation with Sandra Rodriguez tackles the work of Kate Crawford, author of the Counter-Atlas of Artificial Intelligence. This essay explains that “AI systems are ultimately designed to serve dominant interests”. And that it is urgent to ask questions about how they are produced and adopted.
“Kate Crawford said that technology is not in itself good or bad, but that it is never neutral,” says Sandra Rodriguez. We must keep a well-lit antenna that asks questions about artificial intelligence, such as who trains it and on what basis. »
Meaningful decisions
Sociologist Jonathan Roberge, Full Professor Center for Urbanization, Culture and Society at INRS, shares the same concerns.
While he’s not overly concerned about artificial intelligence influencing our decisions on various digital platforms (Netflix or Spotify recommendations, for example), he says that the more “, the more sensitive systems” will be affected,” the more problematic it will be”.
In education, defence or justice, for example.
It doesn’t matter that you don’t watch the best movie one night. But for you to go to prison for eight years following the decision of artificial intelligence when a human would have given you a much lesser sentence… We are not in the same social range.
Are robots about to rule the world 2023 ?
Jonathan Roberge, sociologist and full professor, Center for Urbanization, Culture and Society, INRS
And that’s not science fiction, either. Already, in the United States, some are relying on artificial intelligence for bail decisions. And this new way of proceeding, because of the biases of the algorithms, is far from unanimous. The results obtained in New Jersey have been denounced in particular.
5 Health and Well-being Tips for Building Professionals
The sociologist also believes that we have just taken a step forward in using artificial intelligence for our decisions. “It’s so effective and easy to put into operation in different systems that, through economic pressure, it will spread,” he predicts. Whether it is harmful or interesting for contemporary societies is almost irrelevant. »
The genie came out of the bottle. And he won’t be going back there again.
It must therefore be tamed.
To avoid hitting a wall
So, when will the day when computers will rule the world, when is it?
“I don’t think we’ll ever see that,” replies Catherine Régis, full professor at the Faculty of Law of the University of Montreal.
But developments in artificial intelligence force us to think urgently about how we want to work with machines, she replies.
“We have not yet sufficiently articulated or defined what we want to delegate to the machine and what we want to keep for the human being,” she believes. However, if we don’t do this quickly enough, “we’ll end up in front of a wall”.
His remarks make me think of what Winston Churchill said about change: taking it by the hand before it grabs you by the throat is better.
Speaking of taking it by the hand, both Catherine Régis and Céline Castets-Renard – for her part, a full professor at the Faculty of Law at the University of Ottawa – affirm with confidence that it is high time for our governments to legislate on this subject.
The Montreal Declaration for the responsible development of artificial intelligence, adopted in 2018, caused a stir. But it is simply an ethical framework. It needs to be more.
Are robots about to rule the world 2023 ?
“It’s ethics, and companies don’t care. We need law. I work with very serious ethics professors, and they come to the same conclusion as me,” says Céline Castets-Renard.
A bill on protecting personal information and artificial intelligence (C-27) is currently under consideration in Ottawa. It’s a step in the right direction.
Catherine Régis, Full Professor at the Faculty of Law of the University of Montreal
We must move towards binding frameworks not only at the local level but also at the international level.
Catherine Régis, Full Professor at the Faculty of Law of the University of Montreal
Because the algorithms that will change our lives more every day know no borders, those who feed social networks like Facebook, Twitter and TikTok are proof of this. American and Chinese giants impose their rules on us.
At the same time, what is crucial is to shine our spotlight on all the issues raised.
Are robots about to rule the world 2023 ?
Think about it and talk about it.
Pretty much every expert we interviewed told us that.
Our societies’ lack of serious debate on these issues contrasts with the magnitude of the announced changes and potential abuses.
However, we cannot count on artificial intelligence to take care of it for us!
Defining the orientations of this new Technology is our role, humans.
“I don’t think we need more debates, but we need more conversations that will allow us to explore the tensions between the different properties of systems,” says Joëlle Pineau.
The tensions? She explains that if we develop an artificial intelligence model, we can, for example, make it more efficient, but we risk ending up with more bias. And vice versa. Or emphasize security… at the expense of privacy. Or even the reverse.
Are robots about to rule the world 2023 ?
Choices must therefore be made.
“We have to decide how we are going to build this new socio-technological reality,” she says.
Same story on the side of Yoshua Bengio. While he is very enthusiastic about how artificial intelligence can “accelerate science”, he nevertheless maintains that a “comprehensive reflection” is necessary.
“I think there needs to be a complete overhaul of what kind of society we need, the environmental issues, but also the issues of abuse of technology, of control, of concentrated power through technology,” he says.
And to add, to complete his warning: “Technology can liberate, but it can also be a tool for concentrating the power of wealth. And that is the exact opposite of democracy. »
How to hire remote developers: Everything you need to know
How to start a business? Guide in 13 steps
What you need to know about document management
28 best small business ideas to start in 2023
Are robots about to rule the world 2023 ?