The day that AI became the new villain of the movie
While some debate how to avoid the apocalypse in capital letters, the other extinction, the low-intensity one, has already begun for some time
We have seen it a thousand times, on TV and in the movies: a malevolent actor infiltrates the production and distribution of energy, telecommunications, air or road traffic control, healthcare, or the financial system that makes a city, a country, or the entire planet function, and contaminates or shuts it down until, on the brink of catastrophe, the good guys arrive and neutralize them. The villain, almost always, is a computer genius with monetary aspirations, little social life, and some pending grievance with the established powers, or is in the service of an adversary state. A comfortable archetype that is easy to stop in ninety minutes.
This script, however, has expired. The development of artificial intelligence platforms has reached a point where almost anyone interested, without a secret laboratory or screenwriter clich茅s, can use them to cause chaos: the same tools make it easier to find security loopholes that previously went unnoticed and exploit them in the worst possible way. A malevolent actor is no longer even needed: the normal operation, at maximum speed and without brakes, of the very companies that claim to want to avoid it is enough.
The last few days have made this clear. Jacob Coxon, a researcher who had worked at both OpenAI and Anthropic, resigned last Tuesday accusing both companies of racing toward self-sufficient superintelligence while "gambling with our lives." A few hours later, Evan Hubinger, head of alignment at Anthropic itself, agreed with him in public and estimated the probability of AI ending humanity within a decade at more than 10%. 10% doesn't seem like much, until it's said by your plane's pilot before takeoff.
Technological cynicism
It is unsettling that these confessions of fear arrive just as the two mentioned companies are preparing for multi-billion dollar stock market debuts. That their own employees are coming out to say that the product could exterminate us does not seem like the best campaign to sell shares, unless, like arms manufacturers who boast of the most efficient missile or the deadliest drone, the goal is precisely to exhibit technological power to inflate their appeal to investors. The cynicism is astronomical: if they are so afraid of their products, they should simply turn them off, but with these figures on the horizon, the market weighs more than any ethics committee.
Meanwhile, at the highest levels of the United States government, even more expeditious proposals are circulating. Jacob Stokes, a former White House advisor, has published a report for the think tank CNAS in which he asks Washington to consider cyberattacks, espionage, and even bombings on Chinese artificial intelligence gigafactories, to prevent Beijing from reaching general AI first. The proposal reminds me of the military actions of a few decades ago to free the world from supposed weapons of mass destruction, those for which a former Spanish president claimed to have evidence without blinking an eye. The difference is that it is much harder to show photographs of a malicious algorithm to the UN Security Council.
Someone should stop these tech bros before the damage is irreparable. However, it is worth asking a less comfortable question: if humanity is to go extinct, would it prefer to do so all at once, in a few weeks, due to a lack of drinking water, medicine, or electricity because someone decided to turn off the grid? Or would it prefer to do so slowly, accepting without many questions that the infinite scroll and addictive algorithms of Instagram, TikTok, or YouTube fry the brains of users (not just teenagers), until reaching a society divided between idiotized citizens and a small group of individuals who manage the mechanisms that idiotize them?
Australia and the European Union have already made their move: the former will force platforms to allow the disabling of automatic content recommendations, and the latter uses the Digital Services Act to demand that big tech companies stop designing networks as an addiction machine, something that Washington, in Trumpian times, does not hesitate to label a protectionist attack. A small gesture compared to bombing data centers, but more within anyone's reach than deploying an army. While some debate how to avoid the apocalypse in capital letters, the other extinction, the low-intensity one, has already begun some time ago, and the proposals to stop it are rather timid.
