The past few months have seen a hype about artificial intelligence (AI). In fact, we have spent about five years in which languages have been emerging from companies and academia (especially companies) trained on quantities of human-produced texts for everything one imagines. The starting signal was given, for the general public, by the appearance of ChatGPT. A number of similar instruments have since come out, built by Google, Baidu and others. Many specialize in financial matters, several in logo and image generation. We recently saw artificially intelligent(?) images of Pope Francis wearing a very elegant white coat, and of the capture of Trump.
(You may also be interested in: Putin and his plan B)
At the end of March, an open letter circulated asking the developers to suspend their investigations for six months. It was not directed specifically at anyone. It was something like a prayer to the almighty Lord of the Universe to free us from evil. When I saw it, it already had about 30,000 signatures. The first were from most major software developers: Elon Musk, via Twitter; Steve Wozniak, co-founder of Apple, a number of CEOs of computer companies and, of course, Yuval Noah Harari, who changed his profession from historian to prophet.
The letter had the right components: some doomsday threat, a virtue statement by the signatories, and no practical scope. To begin with, I did not see any Chinese, Russians, Hindus or Iranians among the signers. His style and wording seemed like a product of ChatGPT. As if the signatories had decided to play a joke and had asked the same chat to write a letter warning about its dangers.
We can be scared because it will eliminate routine jobs, or think that it opens the field for many others, perhaps more interesting.
The fact is that two months have passed and the investigation (including the one carried out by several of the signatories) has not been suspended. Moreover, every day something new comes out. It would have been more useful to speed up (no doubt already on the way) the development of a technological tool that allows one or the server to detect AI products. Something like antivirus.
Advertising hype is bad reference to make correct decisions. Keep in mind that this is just beginning. Surprises await us; indeed there are dangers, we have to be alert. But we must recognize that it is an inevitable development; you have to see where the opportunities are and what the real threats are. Without a doubt, it will replace many jobs (it will not be the first time that it happens). It is already evident that he writes better documents than lawyers, diagnoses medical images with great precision and designs work plans, writes reports and does accounting; suggests investment portfolios, and more. But it is very unlikely that he will write one of those magisterial sentences of the great judges, or that he will propose revolutionary economic policies.
It will increase the productivity gap between countries with different levels of development. That suggests us to close the technological gaps; there is no use writing manifestos. We can be scared because it will eliminate routine jobs, or think that it opens the field for many others, perhaps more interesting.
The question of whether artificial intelligence is intelligent does not have a simple answer. But if we ask if you understand the questions we ask, the answer is no. He doesn’t understand, he reviews an immense amount of text, makes statistical associations between words and organizes all this in tables, figures and texts, governed by clear rules. They are not creative or imaginative, they are well compiled; for this reason they are generally coherent, plausible and well written, although sometimes they say nonsense; in that they resemble human intelligence.
MOSES WASSERMAN@mwassermannl
(Read all the columns by Moisés Wasserman in EL TIEMPO, here)
