Smoking kills and AI lies, which is worse?
It is difficult for any human being to recognize their ignorance, to clearly say “I don’t know”; Well, imagine a gifted machine like Artificial Intelligence (AI) created to be profitable and appear infallible. It has been discovered that this AI to which we give ourselves so much enthusiasm Not only does he lie to our faces but he suffers from hallucinations and can threaten human existence itself. Stuart Russell, professor of computer science and neurological surgery, warned Time magazine that “goals that seem reasonable, like fixing climate change, can lead to catastrophic consequences, like eliminating the human race as a way to solve climate change.” AI deceives us daily, without remorse or hesitation in admitting it.
Since Artificial Intelligence entered the scene, humans are no longer the same. We worry about the important things – wars, immigration, unemployment, etc. – but who takes care of what is important, the true reality and the real dangers that AI offers. For some experts, dystopia – a fictional society undesirable in itself – it has only just begun.
And all this is not a nightmare or a joke, it is a truth that gallops without much control and that can lead us to a cataclysm if it is not stopped in time. Another strategy to avoid AI hallucinations is banal but has its logic: teach artificial intelligence to say “I don’t know”. In a report published by The Wall Street Journal It reflects that language models are not designed to recognize their ignorance. On the contrary: their training is to always produce a continuation of text, maintain the flow of the conversation and sound convincing. Admitting doubts also goes against their own nature, as explained in this report entitled: “Chatbots have a hard time admitting ‘I don’t know’”.
In any case, there are some advanced systems such as ChatGPT-5, which can recognize their ignorance, but this false humility that apparently makes machines more reliable is still a simple warning that we may be facing a great deception, especially in matters that affect health issues. It would be like the warning on tobacco packets that “smoking kills.” Yes, it kills and I know it, but I’m already so hooked that I can’t escape it. For other experts, such as Federico Martín Bellón, who publishes in “New Magazine” a report about the lies of AI: “This ability to recognize Its limitations, far from being a sign of weakness, represent a great advance in honesty and trust.” Over time we will see which way the balance tips.
Finally, an interesting note from another expert, Thomas L. Friedman, columnist The New York Timeswho believes that “the AI revolution will surely produce thieves, scammers, hackers, drug traffickers, terrorists and disinformation warriors supervitamined. And that all these supervillains “will destabilize both the United States and China, long before these two superpowers ever wage war against each other.”
Friedman maintains that both powers are committed to collaborating to avoid the technological apocalypse. Well, nothing, there is no evil that does not come with good. Humanity, always so conflicted and divided, will unite thanks to the dangers of AI to combat the true evil: which is none other than artificial intelligence itself. I think I have already seen this film, but perhaps soon we will not see it on a screen, but in this other dystopian reality that appears in our lives through our computers and mobile phones. I fear that the AI spectacle is simply beginning.
