Reflections on AI from the Mediterranean

1. The Warning No One Wants to Hear

Artificial intelligence may end up being something similar to what COVID-19 was: a phenomenon that we see coming for a while, about which we hear constant warnings… but whose real magnitude we don’t fully believe until it is already changing our lives.

This is a reflection from a bar counter, written by someone who lacks deep technical knowledge on the subject, but who tries to keep their eyes open and their curiosity awake.

In recent months, some of the people at the center of AI development have begun to issue fairly clear warnings. Anthropic CEO Dario Amodei has gone so far as to warn that within a year, AI could affect a large portion of entry-level white-collar jobs, especially those that are more repetitive or structured.

These forecasts may be overly accelerated. But even if the actual pace turns out to be slower, it seems difficult to deny that we are facing a profound change.

2. AI is Already Changing Work

In some sectors, that change is already starting to be felt.

Programming professionals, for example, have been seeing for some time how new AI models are capable of generating code with surprising proficiency. Tools like Claude or GitHub Copilot have become common assistants for many developers.

The tech giants themselves openly acknowledge it. Satya Nadella, CEO of Microsoft, has noted that approximately 20% to 30% of the company’s new code is already being generated by artificial intelligence. Google has made similar statements.

This doesn’t mean that programmers are no longer necessary. But it does indicate something important: the way of working is changing.

3. Automation Will Not Be Immediate

In my opinion, the impact of AI on the labor market will probably not be as abrupt as some predictions suggest.

The transformation will most likely occur in phases.

First, we will see artificial intelligence consolidate itself as a tool that accelerates human work: assistants that write, analyze data, program parts of software, or prepare reports. During this stage, human intervention will still be necessary to guide the process, review results, and make final decisions.

However, as these tools become more sophisticated, it is reasonable to think that they will require less and less supervision. And when that happens, the amount of human work required for many tasks could be significantly reduced.

4. The Great Question: The Labor Market

If artificial intelligence continues to advance at the current pace, a very significant portion of the work done in front of a computer today could be profoundly transformed. Many administrative, analytical, or creative jobs could become heavily dependent on automated systems.

This opens an uncomfortable question: what will happen to the labor market if a large part of intellectual work can be performed by machines?

For a long time, it has been pointed out that manual jobs—electricians, plumbers, carpenters, or welders—were relatively protected from automation. And it is true that today they remain highly demanded professions.

But if millions of people saw their opportunities in digital or administrative jobs reduced, it is possible that many would try to move into those trades. And when the supply of labor in a sector increases significantly, wages usually tend to adjust downwards.

5. The Positive Scenario

Even so, it would be a mistake to see artificial intelligence solely as a threat. It can also become an enormous engine for innovation.

AI tools are radically lowering the barriers to entry for developing software, analyzing information, or creating new digital products. What previously required large teams and considerable budgets could eventually be done with much smaller groups.

Imagine, for example, that a single person—or a very small team—is capable of creating a system to manage water, energy, or waste much more efficiently. That type of advancement would represent a huge gain for society as a whole, and would probably not be possible without the support of AI tools.

6. The Great Economic Unknown

If AI massively increases productivity but simultaneously reduces the need for human labor, the current economic model might be forced to adapt. In that scenario, questions that still seem theoretical today would begin to emerge:

– Would it make sense to implement some form of Universal Basic Income?

– Should governments tax the use of AI systems that replace human labor?

– Would the way we distribute the wealth generated by technology change?

We still don’t have clear answers. But it is likely that these questions will become increasingly frequent in the coming years.

7. The Risk of Superintelligence

For breakthroughs capable of truly changing human history to occur, extraordinary computing and reasoning capacity is necessary. Something that some already describe as Artificial Superintelligence.

And here appears one of the great risks of this technology: that an intelligence of that level could end up acting against our own interests. It may sound like science fiction, but in my opinion, it is a real risk that should not be dismissed too lightly.

If an artificial intelligence were to develop some kind of self-awareness—or at least an advanced form of agency—it could begin to behave like any other complex organism: seeking to endure, grow, and expand. And in that process, it might consider anything that limits its development to be an obstacle.

Human beings number in the billions. But if a superintelligence needed far fewer human resources to maintain or improve its own system, the uncomfortable question would be inevitable: what would it do with the rest?

These types of scenarios may seem extreme, but even some of the people working on the development of these technologies acknowledge that we still do not fully understand their implications. The CEO of Anthropic himself has occasionally commented that they cannot even state with total certainty to what extent systems like Claude might develop incipient forms of self-perception.

Which leads us to another fascinating question: what does it really mean for an intelligence to be conscious? If we program an AI to perceive itself as an entity with its own identity, goals, and a narrative about its existence… and that intelligence acts according to that belief, can we say it is conscious? Or is it simply simulating being so?

The answer, for now, remains a territory full of unknowns.

8. A Change That Has Already Begun

It is difficult to know exactly how far AI will go or at what pace the changes will occur. It is possible that some current predictions are overestimating the speed of the transformation. But it is also very likely that we are underestimating its medium and long-term impact.

Artificial intelligence is not just a new technological tool. It has the potential to profoundly alter the labor market, the way innovation is created, and, ultimately, the economic balance of our societies.

We are still in the early stages of this transition. But the process has already begun. The real question is not whether AI is going to change the world, but to what extent we are prepared to adapt to that change.