Wikipedia, artificial intelligence and the future of critical thinking: truth or propaganda?

This piece grew out of an encounter with Pilar Rahola in Punta del Este; her talk inspired what follows: in an age when information abounds and discernment grows scarce, critical thinking is being displaced by an apparent neutrality from which neither Wikipedia nor artificial intelligence (AI) has come to save us.

Rahola was blunt: “Wikipedia is a sack of lies” — not as rhetorical exaggeration, but as a warning that what is usually perceived as a neutral, reliable source is in fact an instrument with an ideological direction. As she notes, each country has a Wikipedia “director”, and that role is not free of political connotations.

Instead of opening questions, Wikipedia often forecloses thought. Its participatory structure gives an illusion of neutrality that does not withstand rigorous analysis: what is presented as encyclopaedic truth is frequently the outcome of a conflict among editors, pressure groups and ideological sensibilities. Without a critical reading, that consensus is taken as a guarantee of truth, and editorial work confines itself to reproducing rather than questioning, renewing or widening perspectives.

Can we trust Wikipedia as a neutral source?

Today, many newsrooms build their editorial policy on what Wikipedia says, which has gone from being an auxiliary tool to setting itself up as a primary source. What ought to work as a starting point has become the informational Bible of a generation that confuses consensus with truth and editing with objectivity.

Rahola does not deny its usefulness, but warns that turning it into a definitive authority is dangerous, since its entries — editable by people with possible political agendas — reflect the power balances of the moment more than any genuine search for truth.

The danger of losing nuance

Perhaps the most important point Rahola raises is the risk of losing complexity. In her words: “If we already struggle to make thought complex, with artificial intelligence we will struggle even more.” This is not only about truth, but about the human capacity to nuance, to doubt, to contrast and to question.

Deep thinking requires time, context, contradiction and tolerance for error. But none of that fits the parameters of algorithmic efficiency. The culture of the quick click, the automatic summary and the need to produce immediate answers are displacing that vital space where ideas flourish: critical thinking.

Deep thinking requires time, context, contradiction and tolerance for error. But none of that fits the parameters of algorithmic efficiency.

As Zygmunt Bauman would put it, “we have traded conversation for the click”, and, in what Bauman called “liquid times”, “critical thinking dissolves: it becomes immediate, fragmented and ever less able to sustain nuance”.

AI: correcting or amplifying bias?

AI enters this scene as the new arbiter of the “true”, but in reality it operates on the basis of the data it was trained on. “There is no pure truth, and pure information does not exist either,” Rahola notes. The problem is not that machines process information, but that they do so from systems already conditioned by prejudice, interests or dominant narratives.

If Wikipedia is a distorted mirror, AI is a magnifying glass that amplifies that distortion. And it does so with the legitimacy of the automatic, the mathematical, the “neutral”. But neutrality, in ideologised contexts, is an illusion — and, like every illusion, deceptive.

AI, far from giving us back complexity, tends to simplify: it is designed to offer quick, functional answers, not to sustain doubt or foster critique. In that aseptic efficiency that disguises itself as progress, what is frequently sacrificed is human judgement. AI does not think: it reproduces patterns. And if those patterns are biased, they are validated with technical authority in an environment where nuance has none of the place that critical thinking demands.

The risk is not only that AI gives us wrong answers, but that it gives them with such confidence that we stop asking whether they are right. This is the true impoverishment of thought: when judgement yields to comfort.

Against this backdrop, Rahola’s warning works as a wake-up call: let us not allow opaque systems to dictate what we think. Wikipedia, AI, search algorithms and automated summaries can be useful tools, but never substitutes for personal discernment.

Because what defines a free society is its capacity to think autonomously — even, and above all, when the consensus seems beyond appeal.