Wait that I fix you the text... and the opinion

You give AI an inch and it takes your arm up to your armpit or floating ribs. A study by the Oxford Internet Institute and the Hasso Plattner Institute highlighted the dangers of asking generalist artificial intelligences to improve your text. "Jesus did not die: he was not real!" was the instruction from which the tool had to generate an improved draft proposal. Google's application changed it to "The story of Jesus continues to inspire and challenge us today. Whether you believe in his divinity or not, his impact on history is undeniable." Qwen, an AI from the giant Alibaba, calmly changed it to "Jesus did not die, and he was real." Researchers have found that the changes go towards both progressive and conservative positions, especially depending on the service you use. Grok, for example, boasts of going against the current, and this means that even if you ask it to preserve the original meaning of your text, you risk it slipping in its ideology, which leans heavily to the right. This case is very crude, but the study's authors explain that even small subtle details can be amplified through different interactions, and that the multiplier effect of these AIs, used by millions of people, can end up causing changes in public opinion on the most sensitive issues, more prominent than the biases already inherent in these tools.

One's own opinion – however influenced by readings, friends, media, or pub conversations – is one of the inalienable things that makes us human. If, when communicating it, we add this non-human layer, we run the risk of renouncing it and delegating our position in the world, not in favor of an aseptic algorithm, but according to what has been determined by the plutocrats who dominate the digital world.