The possibility offered to the public to use ChatGPT will have had the immense merit of placing artificial intelligence (AI) at the heart of our debates. Finally we talk about it, even if in a messy way. The configuration of the public debate, very divided, we are asked to choose our side: either on the side of Jean-Noël Barrot, our Minister Delegate in charge of the Digital Transition (ah the titles of ministerial posts…) who considers ChatGPT as an “approximate parrot “, or on the side of the Laurent Alexandre who see in it a major economic, social and civilizational change. Of these two extremes, it seems to me that the second is the closest to reality.
ChatGPT is a so-called generative artificial intelligence application. This means that she generates texts or lines when talking to humans. Other applications can generate images, videos or music. ChatGPT is one avatar among others of the third industrial revolution, which sees the convergence of digital technologies, artificial intelligence, robots and biotechnologies. Put in this perspective, ChatGPT is one application among others, most of which do not yet exist. The future is full of surprises.
To our Minister Jean-Noël Barrot, whom we knew better inspired than at the time of this ornithological projection, we will point out that, not only is his judgment severe, but above all ChatGPT will be 100,000 or 1 million times more powerful in ten years, artificial intelligence learning exponentially thanks to the data provided to it and to its own experience. It is to this forward-looking effort that we must compel ourselves to make the right public policy decisions.
A broader reflection on ChatGPT
At the end of February, Henry Kissinger (the diplomat), Eric Schmidt (the former CEO of Google) and computer science professor Daniel Huttenlocher published in the wall street journal a paper too little commented on in Europe. Their diagnosis is as follows. The consequences of generative AI applications go beyond classical Schumpeterian “creative destruction”. We are obviously right to wonder what these apps are changing in terms of jobs and the way of working. But our reflection must be much broader. First, write the three experts, ChatGPT changes our relationship to knowledge, more than printing. We know in principle the identity of the author of a book and we must be able to get an idea, even imprecise, of his skills. On the other hand, what is the veracity of the statements generated by an AI? This is a good question to discuss in philosophy class.
Next, generative AI raises the question of technological and economic power. Because the artificial intelligence sector is highly capital intensive. This means that these markets are naturally oligopolistic. Those who invest the most, soonest and best will be the generative AI giants, that is, they will hold major technological and economic power. Today, it is obviously American and Chinese organizations that dominate. Finally, geopolitical reflection must encompass the issues of generative AI. Look at the presence of Russian trolls on social networks, Facebook and Twitter in particular. Imagine what Russian influence would be if a generative AI could develop a credible (albeit false) argument that Western economic sanctions do not affect Russia but ultimately the West itself. Geopolitical power requires mastery of AI.
ChatGPT should also be an opportunity to take a critical look at our public debate and the state of opinion. Without any contempt for anyone, we can still consider that the crisis of national hysteria around an essential and minimalist pension reform is a little off the mark even though the underestimation of technological change threatens France with relentless downgrading. Let the law be quickly promulgated so that we can talk about important and interesting things again.