Scientists find artificial intelligence weakens human intelligence and spreads like a virus

An international research team from Spain, Italy and the United States published Workwhich compared large-scale artificial intelligence language models to cognitive viruses. Not only can a model impair a person’s ability to think, it can also have a self-reinforcing effect if those around them actively use it.

Photo credit: Steve A Johnson / unsplash.com

The work is theoretical rather than experimental, but also addresses previous evidence on the impact of AI addiction on cognitive function. Using epidemiological modeling techniques used to track the spread of biological viruses, the scientists concluded that societies shift from one state to another over time: “independence” first becomes “attachment” and then “dependence.” In the first case, people do not rely on AI models for their thinking processes; in the second, they use them as part of their cognitive processes; and in the third, they have transferred most of their mental work to AI.

The study’s authors aren’t limited to claims that artificial intelligence is damaging the brain — previous research has also shown this. When a person turns to a machine to perform mental tasks for him, his ability to think independently is diminished. A new study suggests that widespread adoption of artificial intelligence models could have systemic consequences for society as a whole. The ongoing practice of delegating psychological functions to artificial intelligence may act like a biological virus: People who benefit from artificial intelligence work spread the practice to others and, unless they exercise due diligence, gradually become dependent on it.

The authors of the study are not opponents of artificial intelligence; rather, they emphasize that they do not believe that the use of artificial intelligence is inherently harmful and give several examples of how artificial intelligence can actually help the human mind without replacing it. That’s the key to immunity to viruses—not letting artificial intelligence think for itself. There’s no need to give it up – models excel at deep analysis, information gathering and even training tasks. The question is not whether to use artificial intelligence, but how to interact with it: what matters is whether it expands human capabilities or replaces the mental processes that support such capabilities.

The authors cite research showing that thinking critically about artificial intelligence without sacrificing your own mental work is key to preventing mental decline and a way to build immunity to viruses. It’s important to stress that this work is not an experimental study – scientists arrived at their conclusions by applying complex mathematical instruments to existing scientific data.