Harari and the Fear of Artificial Intelligence

In the fictional universe of the Dune books, it’s strange to realize that artificial intelligence doesn’t exist. Despite the very distant future conspiracy in which humanity has occupied the entire universe, no artificial intelligence is in sight. In the distant future of Dune, humanity has already created artificial intelligence. But they were so dangerous and destructive that in those distant times they were completely banned.

Last week, historian Yuval Harari and designer Tristan Harris published an article on artificial intelligence that evokes feelings similar to the Dune saga. From their point of view, the artificial intelligence that is now becoming en masse in the form of products like ChatGPT, MidJourney and others is an “alien” force that is being invoked without us even being able to understand how it works.

The main competence of these artificial intelligences is nothing less than hacking the pillar of civilization: language. It is through language that we relate to one another, build institutions, develop laws and culture, declare and end wars, and coordinate the actions of individuals and groups. Well, now there’s a machine among us with the potential to speak the language better than anyone could ever aspire to.

In Harari’s view, by allowing artificial intelligence to talk to us and learn from us, we are handing the key to the civilization process to the machine. It’s worth noting that the most widely held definition of artificial intelligence is that they are “intelligent agents that perceive their surroundings and make decisions that maximize their chances of success.” The most important word in the definition is the term “agents”, i.e. tools that are supposed to act on our behalf.

But what if these agents start acting against our interests? In Harari’s words: “Democracy, for example, is a conversation. Conversations are conducted through language. When the language itself is hacked, conversation breaks down, democracy becomes untenable. If we wait for chaos to act, it will be too late to do too much fix”.

It is a fact that we allow artificial intelligence capable of talking to us to become accessible to hundreds of millions of people without thinking deeper about its implications. And more importantly, without taking safeguards that can mitigate their effects, many of which are unpredictable.

Harari’s article circulated quickly because of its alarmism. But there is another, more interesting and analytical text worth reading. Its author is the mathematician and computer scientist Stephen Wolfram. Its title asks, “Will AIs take away all our jobs and end human history?”. Wolfram gives the following answer: “It’s complicated”. In the text he shows that it is no longer possible to think that there will be human activities that cannot be replicated or replaced by artificial intelligence.

Given this, Harari suggests that institutions quickly develop to contain and then govern artificial intelligences. Wolfram suggests that our challenge is to spread computational thinking so that we can understand it and act accordingly. The truth is we have a new problem that is still looking for an answer.

it’s over no internet laws

Already the wave of legislative proposals to regulate internet platforms

It’s coming the wave of proposed legislation regulating artificial intelligence