Artificial Intelligence: Revolution or Bubble?

The Storm Reshaping Science and the Global Economy

In 2014, many economists viewed Artificial Intelligence (AI) as the engine of a new age of hyper-productivity. A decade later, the debate is still open: fears of an economic bubble are rising while enthusiasm builds around a scientific and social revolution without precedent.

As AI analyst Cade Metz notes, every major technological shift has sparked its own wave of anxiety: lost jobs, disrupted companies, entire industries turned upside down. The key question, then as now, is what will emerge once the dust has settled.

In addition to market turbulence, AI is already rewriting the rules of scientific research. Its ability to process massive, complex datasets at speeds impossible for humans has turned it into a true engine for discovery. Drug development offers the clearest example.

Traditionally, bringing a new drug to the market required more than 10 years and roughly $2.5 billion in costs. Today, AI can slash both time and cost by sifting through millions of genomic, proteomic and clinical data points to pinpoint molecules or biological pathways linked to disease. Machine learning systems can generate and simulate millions of compounds in silico, predicting efficacy, toxicity and target affinity—a high-speed computational screening that condenses years of lab work into a matter of months.

These models can also optimize molecular structures to improve key pharmacokinetic properties (ADME: Absorption, Distribution, Metabolism, Excretion). This synergy has already led to breakthrough discoveries, including new antibiotics against resistant bacteria and promising therapies for several cancers.

In hospitals, AI tools analyze brain CT scans and alert neurologists to early signs of stroke, cutting intervention times by crucial minutes: often the difference between recovery and permanent disability. In breast cancer screening, studies on a population of tens of thousands of women show that AI can detect more tumors than radiologists, without increasing false positives.

Yet despite AI’s power, human judgment remains at the center of research. AI computes but it does not reason. It can test billions of hypotheses, but it is the scientist who formulates the initial intuition, sets the goal and asks the right questions. The “logical leap” that links two distant biochemical pathways—a leap from which major discoveries arise—cannot be automated.

In addition, AI does not possess a moral compass. It can optimize a drug based purely on efficacy while overlooking long-term side effects unless human-defined scientific and ethical safeguards are in place. It is always humans deciding what problems deserve attention and how results should be used.

Interpreting data is also an irreducibly human task. When a model produces an unexpected output, only experience can distinguish a glitch from a potential breakthrough. AI is the microscope of the 21st century: it sharpens our vision, but the eyes and the mind remain human.

Economically, the current AI boom raises tough questions. Are we heading toward a bubble? Tech giants like Google, Meta, Amazon and Microsoft already generate vast profits, but their valuations are further inflated by their near monopoly over AI infrastructure, especially cloud computing. Thousands of startups entering the field must purchase compute power from them, thereby fueling Big Tech’s revenues even more.

But what if these startups, weighed down by skyrocketing costs, fail to turn a profit soon? The whole system could shake.

A burst like the dot-com crash would hurt investors and wipe out many companies but would not likely threat the global economy. The deeper concern is a scenario closer to 2008: some investors are heavily leveraged, and a sudden downturn could trigger a broader crisis of confidence and liquidity.

One question keeps resurfacing: will AI replace doctors? It’s the wrong question. AI has no patients. It doesn’t wake at 3 a.m. worrying whether it missed a diagnosis. It doesn’t look a family in the eye when there is nothing left to do. It doesn’t carry the weight of those words through the rest of the day. The real question is whether we will remain human while using increasingly powerful machines. Whether we will remember that behind every dataset lies a person, behind every algorithm a choice, behind every optimization a responsibility.

AI computes, we decide. And decision-making—rooted in the responsibility to care for others—remains an exclusively human prerogative.

Bubble or revolution, one thing is clear: AI has opened doors to scientific advances that only recently seemed unimaginable. The real challenge extends beyond markets as it lies in shaping a new generation of researchers able to pair the power of algorithms with the creative, critical and irreplaceable intuition of the human mind.