AI labels Spanish forum Burbuja.info as far right

Artificial intelligence and the far-right label: reflection or bias?

A user asked Google AI whether a well-known Spanish online community leaned ideologically towards the far right. The answer was yes, citing academic research and media outlets as sources. The news spread like wildfire among the community's members, who reacted with scepticism, irony and, in some cases, resignation. The debate that trinc questions not only the label itself, but also the reliability of AI systems when classifying ideologies in digital spaces.

The AI's answer: uncomfortable truth or self-fulfilling prophecy?

The original prompt asked whether the forum leaned to the far right. The AI, based on a language model trained on large volumes of text, said yes, backing its answer with references to studies on the "populist radical right" and reports from outlets such as El Salto Diario. The answer included terms like "gender violence denial" and "misogyny", which set off alarm bells among users, who see those accusations as a wildcard used to discredit any criticism of the status quo.

One line of analysis holds that the AI merely replicates the bias of its training data, drawn mostly from conventional media and pogre-leaning academic publications. For them, the classification reflects the dominant editorial line, not an objective diagnosis. Others, however, argue that the forum's content — with frequent references to conspiracies, attacks on the left and acid humour aimed at "political correctness" — fits perfectly into what analysts consider the digital far right.

Who defines the far right on the internet?

The debate raises a deeper question: who has the authority to label ideologies online? AI critics point out that the categories are imposed by an academic and media elite that pathologises all dissent. "Everything that isn't proge is far right", one participant sums up ironically. The AI's answer, citing ResearchGate studies and reports from El Salto Diario, reinforces the idea that there is a knowledge network labelling as extremist any voice outside the mainstream consensus.

On the other hand, those who defend the classification argue that denying the evidence is a typical strategy of extremists themselves. They point out that the forum hosts multiple threads and users with diverse opinions, but that the dominant stance is clearly radical right, and that members themselves know it even while denying it. "Fascists are the vast majority on the forum", says a user who describes himself as non-aligned.

The AI dilemma: to reflect or to judge

The controversy reveals a dilemma inherent to language models: they do not distinguish between describing and prescribing. When the AI says a space is far right, it is not offering a neutral observation but participating in the construction of that reality. A label issued by a system many consider objective carries a weight that can stigmatise the community and its members.

---

The open question is whether artificial intelligence can be a useful tool for identifying extremism or whether, on the contrary, it becomes a weapon of ideological censorship. Meanwhile, the community keeps debating: some with the calm of those who know labels do not define their identity, others with the anger of those who feel unfairly singled out. The only certainty is that AI, like its creators, is never innocent.

Related debates on the forum

• Tudela withdraws the winning Santa Ana 2026 poster after its creator admitted using artificial intelligence · 53 replies
• Do you consider yourselves white? Have you experienced racism in Europe's white countries because of your looks? · 191 replies
• Sheryl Crow like a mummy dragging herself across a stage. She's completely finished and doesn't realise it · 50 replies
• An idea repeated a thousand times - it's very easy for me to understand why there was a civil war in this country · 43 replies