I was writing a thread on Twitter to express my reflections on applying a moratorium on the development of artificial intelligence, a symbol of how the way we communicate has already changed in our days, when I realized that I preferred to write a longer text and restful, a symbol of how difficult it is to adapt to changes for a semi-intelligent middle-aged mammal. And adaptation is precisely what I want to talk about. And also ethical and moral issues. And yes, I want to reflect because I am far from having forceful and indubitable answers in each of the aspects of the debate, although I do have some certainties.
In recent days, the manifestos and statements by some experts and public figures have been repeated, some as controversial from the point of view of ethical honesty as Elon Musk, advocating a moratorium on the research and development of artificial intelligence ( AI), due to the enormous risks that it entails and the high probability that at some point it will escape “our control”. Like Saul arriving in Damascus, a sudden light has been turned on in their heads so that, after decades of AI research, they now realize that a scenario in which, as was the case in Terminator, Skynet decides that humans are a threat to planetary sustainability is more than just a good argument for a dystopian fantasy. And I would say that they are not wrong in their diagnosis and neither in their fears, but I do believe that they are wrong in that neo-Luddite drive that inspires them as a conclusion. Among other things, because history shows us that human beings are perfectly capable of ruining our present and future without the need for the assistance of any cybernetic brain. The red button can still be pressed with one finger. Kubrick showed us that the world could end at the pace of a Texas rodeo.
It is true that the irruption of a generative artificial intelligence such as ChatGPT has put into the public space a debate that was taking place in very small areas, and that is eminently good and necessary, but perhaps the threatening bias clickbait of much of the news generated is not the best way to do pedagogy. I have been working on a European AI regulation since 2019 and I am acutely aware of the enormous risks, but also that most of them have to do with misuse or malicious use by some human being or group of them. It is true that this technology has the unique and novel capacity of being able to evolve by itself without the assistance of a human and without, at times, being able to anticipate exactly in what sense it will do so. And all at a speed that is increasingly dizzying. This poses challenges for us as the adaptive animal that we are. Do we have the capacity to adapt at such speed to technological advances? This must be answered from anthropology and, in any case, study remedies if the answer is not positive —I take this opportunity to, in light of the last statement, assert that digital illustration and the training of experts cannot be reduced to the field of computing or of the exact sciences: before the civilizational change that is coming to light, we need the help of philosophers, jurists, artists…—.
Before, I was talking about being aware of the risks, but I am also aware of the astonishing leap in well-being that this technology can bring us. A few days ago I was with a cardiologist who works with AI models for diagnosis and he was excited about the patterns they are discovering through the analysis of a simple heartbeat to diagnose diseases with astonishing anticipation, establishing relationships with ailments that no one would have. could even suspect. In the field of world productivity, the most conservative studies say that the massive incorporation of AI will multiply it by seven; well it is true that this is not per se good news for the majority of humanity, it all depends on our ability to fairly distribute its benefits.
The European institutions have been working for years on the establishment of a legal framework for artificial intelligence. The European Parliament prepared a first legislative proposal on ethics in AI (for which I was a speaker in 2020) and today it is finishing the discussion on a holistic regulation that will be approved, in all probability, under the next Spanish presidency. Europe is ahead of the rest of the world with the intention not only of finding a virtuous balance between development and protection of our citizenship rights, but that this serves as a compass for the establishment of a shared global scheme. The call Brussels effect It logically establishes incentives for international agents to adapt their operations to European parameters, if they want to work in a rich market of almost 500 million people. And this I think is the way.
I understand that the alarm is raised and it is time that much-needed debates about the civilization change that AI will bring us are addressed publicly. A democratic society has the right to govern that change, establishing rules and limits. But I don’t think a moratorium on development is useful or realistic. The world is much bigger than the West – we should repeat this to ourselves every day as a kind of moment of death, because its non-observance only causes us problems in understanding reality—and this hypothetical moratorium will not occur in all countries, not even in the majority, which would only abound in a development gap that we cannot afford. In terms of AI, months means years in other technologies, let’s think about the expert Nuria Oliver recently stated that there are scientific studies signed at the end of 2022 that are now obsolete.
Above I spoke of the general pedagogy that our societies deserve, the news should underline many more things that Frankenstein’s monster is on the loose. And he was also referring to the tremendous benefits that AI can bring. Stopping research and investing in AI also entails a moral dilemma in terms of solving the multitude of penalties that humanity drags. Let’s think about the fight against diseases or against climate change, in multiplying the efficiency of crops or in the efficient management of water use… Stopping now also has that cost.
I have doubts about a proposal that I consider unrealistic. I don’t know of a single case in the history of mankind in which a technological or scientific development has come to an abrupt halt, or, even less, has been reversed. I think it is time to democratize reflection and set limits on the uses and control in the implementation of the models, as the European institutions are doing. And that this debate is also on a global scale. And also to be aware that we need huge investments in human training, control instruments and governance structures, on the scale that this technological development places us.