Sánchez launches 'Hodio' to track online hate speech without defining it

Hodio: Measuring Online Hate Without a Definition

19 seconds. That is the duration, according to one user, of the video announcement for 'Hodio', a tool to systematically measure hate speech on digital platforms. Less time than it takes to read this paragraph. With this material, the issue has already split into two camps: those who see a thought police disguised as a scarecrow, and those who have long demanded data on what is posted in comments.

The striking aspect is not the reaction, but the void. No one knows exactly what the tool measures, by what criteria it classifies content, who develops it, or what happens with the results. The announcement arrived without a user manual, and that—more than the tool itself—is what has ignited the debate.

What is 'Hodio' and the project's current status

The official approach is limited: a system to measure hate speech circulating on platforms in a 'systematic' way. No sanctions, no fines, and no administrative proceedings appear in the announcement. Only measurement. The key word is systematic, implying methodology, homogeneous criteria, and a comparable scale over time.

From there, speculation takes over. It is suggested that the model might be fed by a database of texts accumulated since 2006, and the Spanish case is compared with the UK, where something similar reportedly works for two years. Neither of these claims is accompanied by an official document. They are hints, not data.

Who decides what constitutes hate speech?

Here lies the crux. Without a closed definition of what falls under this category, the tool will measure something, but no one will know what. The question reappears in various forms: who sets the threshold, what happens with harsh political criticism, and how to distinguish uncomfortable opinion from insult.

One school of thought argues that the problem is not the meter, but the catalog of punishable acts, warning of a chilling effect: people stop writing debatable things not because they are illegal, but out of caution. The counterargument has its own merit. Without public data, the true extent of hate speech is debated through impressions, which benefits those who deny its existence. Measuring, even with an imperfect instrument, is defended as better than not measuring.

The detail of the 'h' in the name also fuels skepticism: if the label is constructed with irony, it is hard to believe the methodology will be surgical.

Parliamentary support and the most repeated statistic

There is a statistic circulating amidst the noise: only one parliamentary group reportedly voted against the initiative, while the main opposition party supported it along with other forces. The information comes from a publication shared with the announcement, not an official document, and should be treated as unconfirmed data.

If sustained, it leaves an uncomfortable reading for those trying to present the tool as an exclusive idea of the Executive. Because the underlying debate is not whether hate speech is measured, but who signs off on the measurement. And there, the coincidence between political blocs that hardly agree on anything disconcerts more than a few.

Another subcommittee and another technological gimmick?

The most repeated suspicion is not ideological, but budgetary. It points to the creation of a trinc-up subcommittee—with corresponding allowances—and the possibility that development will be subcontracted to a public tech firm or handled by civil servants. From there to the memory of other technological launches that almost no one talks about is a step taken without shame.

The criticism mixes verifiable parts with unverified ones. Verifiable: in Spain, each announcement of this kind carries doubt about the cost and who gets the contract. Unverified, for now, is everything else. The project has just been born, and no budget or award has been published. The million-dollar question—literally—is which company or administration will end up assigned the contract.

From irony to prison sentences

The tone rises in stages. It starts with easy jokes and continues with the prediction that there would be no prisons or courts sufficient for so many convicted; in the thread, people talk of more than 10 million people, a figure no one calculates with data but which is repeated with gusto. Inevitable references to 1984 and various dystopias appear, already becoming a literary genre of their own for each regulatory announcement.

In the final stage, the conversation bifurcates. On one side, trench humor: how to recognize each other among inmates if things get worse, with a signal manual included. On the other, a much less funny drift, with messages sliding into disqualifications against specific groups—migrants, LGTBI people, independentists—and some calls to resolve the issue through violent means. It is the minority and noisy part of any conversation about censorship. It should be read as a symptom, not a program.

---

Preventive censorship or basic statistics? The announcement lasts 19 seconds. The response will last much longer, and it will likely be written by an award that does not yet exist.