Four months after the launch of ChatGPT – the conversational agent of the American company OpenAI – Italy banned it as well as, in bulk, Sciences Po, the city of Montpellier or New York schools. 1,100 tech luminaries and thinkers, including Elon Musk and Yuval Noah Harari, have also called for a six-month moratorium to halt the rise of AI.
A moratorium on AI: it’s totally ridiculous, hilarious or horribly cynical. We will gladly go for the last option with regard to the identity of the signatories of the open letter. The very people who have invested billions of dollars in AI and are fiercely competing for control of it will now wisely stop to think about the big blunders they have done. Of course. There can be no moratorium on AI. Nor can there be a moratorium on genetics or nuclear. And no more than yesterday there could be a moratorium on the use of fire. The search and progress dynamic does not have a pause button.
Not only can there not be a moratorium on AI, but that would not even be desirable. Because the benefits of AI are already immense. We are no longer aware of it, but it is already everywhere around us. And makes our life easier. It is an AI that allows us to access all the knowledge in the world in one click or to have an unparalleled radiological diagnosis.
Fake reviews
None of the criticisms made today against ChatGPT and its competitors, since they are the ones who are specifically in the hot seat, hold water. Taking over the world? We are not in terminator. The AIs we have have no consciousness and no will of their own. Use of personal data? This is exactly what all the Gafams (Google, Amazon, Facebook, Airbnb, Microsoft) are already doing. Americanization and standardization of thought? LOL, as the young people say. We have been under American cultural domination for seventy years. And the single thought has imposed itself since the fall of the Berlin wall. Criminal tool? It’s true, ChatGPT can be hijacked by hackers. In the same way that one can use a car to make a robbery or find on the Internet judicious advice to poison his neighbor. Job threat? It is completely unclear what impact this type of AI may have on employment, the most likely being a mutation rather than a destruction. Mental manipulation? A person was allegedly incited to suicide by ChatGPT. But, for a fragile person, living on the sixth floor is also an incitement to suicide…
We could continue the list for a long time. What it bears witness to is the fear that great inventions have always inspired. Yes, the current outcry against ChatGPT is neo-Luddism, the name of this popular movement which, at the beginning of the 19th century, destroyed mechanical looms. A debate which unfortunately illustrates in a caricatural way the joke according to which when someone invents something the Americans make a business of it, the Chinese copy it and the Europeans regulate it. We only know how to do that. Luckily not everyone. While Italy banned ChatGPT, the ECB and French business schools decided to authorize its use in their respective competitions from 2024. While some put themselves under glass, others get on the train.
The precedent of printing
And that’s what you have to do: get on the innovation train. Accelerate on the AI, not slow down by believing in hypothetical moratoriums. Accelerate from a technological point of view: if we consider that ChatGPT is not European enough, nothing prevents us from developing a competing model. More broadly, the race for AI is on. We are still far from having a strong AI, that is to say qualitatively surpassing the performance of the human brain. However, such an AI could open dizzying perspectives in terms of health or ecology.
Accelerate also in the field of the understanding of the tool. Tomorrow, we will have to work with ChatGPT-type AIs. Tomorrow, working may even consist of knowing how to use it. That is to say, knowing how to talk to him, in other words how to formulate the right requests. And know how to process or reprocess its rendering. Yes, Gaspard Koenig is right to point out that ChatGPT texts are not sourced. He sees in it a danger for knowledge and even a major risk of confusion between true and false. At the time of “deep fake” and other “fake news”, the debate is not new. Learning to live with AI, and even all new technologies, means learning to remain critical of them.
The same criticisms were made when printing was invented. Considering that the world was not ready for such an invention, Francis I even banned it. Understandably, printing was an unprecedented means of spreading anything, knowledge or lies. Today’s world is no more ready to welcome ChatGPT than that of Francis I to welcome the printing press. But he adapted to it and who thinks today that Gutenberg’s invention should be banned? Embrace progress. Let’s learn how to use it. And let us rejoice in such a time.
I specify that this text was not generated by ChatGPT.
* Sustainable development adviser to the Senate, Antoine Buéno recently published The collapse of the world will probably not take place (Flammarion).