Will We Still Be Working Tomorrow?

"According to the International Monetary Fund, artificial intelligence will impact 60% of jobs in economically advanced countries, and only 26% in low-income countries"

"The jobs that will be protected are the highly and lowly qualified ones"

"There won't be enough work left, we need to implement a universal basic income"

These are examples of statements you see everywhere right now about AI, and opinions are polarized to the extremes: either the enthusiasm around AI is at its peak and people only see the positives, or they point to the impacts that will lead to catastrophe.

Clearly, this revolution will have a significant impact, and it's very difficult to know what will actually happen in the coming years.

Will we live in a society where we have "robot slaves" that handle what's needed to keep our society running? We'd be left to philosophize, manage political and social challenges, and perhaps set off to conquer new galaxies.

Will we live in a society where there isn't enough work for everyone, requiring us to rethink our relationship with employment? Just as the industrial revolution enabled us to work less, we may work even less with 3-day weeks and 5-hour days.

Will we live in a society where we have the choice of having the body of robots but the consciousness of a human, perhaps even becoming cyborgs?

Will we live in a society where it's impossible to tell a human from a robot, even with a Turing test, like in the TV series Westworld?

Will we be at war with robots like in Terminator because AIs have decided that our human species is a poison to humanity?

Will robots treat humans as an energy source to exploit, as in The Matrix?

All of these scenarios have already been imagined in various works of fiction and, of course, they are also part of our collective imagination.

What is true is that AIs are everywhere and have been part of our daily lives for years, invisibly.

But since the advent of ChatGPT, something new has emerged: the democratization of this technology on a planetary scale, accessible to everyone, and that changes everything. Since then, a frantic race has been launched. Every day we see new developments, and the pace is incredible - and that's precisely what makes this subject so fascinating to us.

How We Think About AI Today

Visible and Invisible AI

I'd like to draw a distinction between what I call visible AI and invisible AI.

Visible AIs are the ones we see explicitly: Alexa, Midjourney, ChatGPT... These are AIs where we all know we're facing an AI and we choose to interact with them.

Invisible AIs are those where we don't know they're there, or when we do, it's rather vague. Here's an example that's happened to everyone: I'm talking with my wife about wanting to go skiing, I pick up my smartphone, open a social media app, and it "coincidentally" shows me ski vacation ads. I recommend watching the excellent documentary "The Social Dilemma" if you want to dive deeper into this topic. It's fascinating.

AI at Work

In 2016, when Crédit Mutuel deployed IBM's Watson AI technology to help and assist bank advisors, there was a major backlash even though the goal was clearly to help advisors work better - which is exactly what happened.

When Cédric Villani published his report in 2018 on AI, a report I still find just as relevant today, he envisioned AI as an assistant to humans.

For months now, I've been using AI almost daily for all sorts of things: it allows me to go faster and further in my work. Yet it doesn't replace the critical eye that I alone must bring, and the judgment that only I can provide.

All of this to say that, as of today, visible AIs are super assistants that are clearly an incredible help for doing things better and faster.

When ChatGPT made a massive BOOM, at Reboot, a company where we develop with programming languages, we had all sorts of debates: as a developer, will I be replaced? What impact will this have on my work?

When you see the arrival of Microsoft Copilot, an AI capable of producing code, it can be scary for developers. In reality, on the ground, what we observe is that Copilot is clearly helpful, but you still need a critical eye to fix what doesn't work and integrate it into a broader context. In fact, we believe that when a developer doesn't have enough development experience, they should not use Copilot because it's an essential phase that can't be skipped: you need to understand what you're doing, how you're doing it, and why. Without that, the developer will take everything Copilot produces at face value, which is very bad and leads to problems.

In other words, once again: AI helps but doesn't replace.

At least not today...

What About Tomorrow?

In March, we have a Let's Talk (webinar) that will cover this subject in more detail, and I don't want to spoil too much of what Tristan will present.

BUT

On a personal level, I can say that to face these new challenges, it's important to educate ourselves as much as possible on these topics. When the questions arise concretely for all of society - because for now, the impacts are still contained - we will need to take a stance and show creativity in designing a new kind of society.

And that is something AI will never be able to do for us!