As we have documented before, the application of community rules has many gray areas and high margins of error. In particular, restrictions related to specific words often lead to wrong decisions. Two recent cases in Colombia show the risks to which users are exposed not because of their content, but because of this assessment devoid of context.
This happened to user Alegandro -@salgadolfo - who on March 13 was banned for twelve hours on Twitter for a comment to a tweet by Argentine political scientist Agustín Laje, who had posted a video on the financing of vice-presidential candidate Francia Márquez. In response, Alegandro claimed that Laje's opinions were biased and that he wanted to be the "Ben Shapiro sudaca", alluding to a US political commentator of the same ideological line as Laje.
Shortly thereafter, Twitter informed the user that he had been sanctioned for violating its hate speech policy. This rule prohibits, among other things, directing slurs and epithets to demean or reinforce stereotypes based on racial, ethnic or national origin, sexual orientation or religious affiliation.
The use of the word "sudaca", a derogatory term to refer to South American immigrants in Spain, triggered the sanction. Platforms often flag certain terms to detect prohibited content. However, it takes much more than the use of a word to know if there is, in fact, hate speech. And while human moderation teams try to add that context into the decision, errors abound.
Some social media have exceptions to the use of insults against a minority and allow, for example, that they be used to raise awareness, to reject them, or as a form of empowerment.
In its hate rule, Twitter allows that "members of a protected category may refer to each other using terms that would not normally be considered an insult". In this case, let us recall, the sanction was applied to a South American user who was precisely addressing another South American user.
However, the tweet in response to Agustín Laje was deleted and the user's account was suspended for twelve hours. On the platform, the content was marked with a tag explaining that the community rules had been violated.
A similar situation arose for the tweeter Robertico -@MarioRobertoP- on June 16. It all arose because someone had published a false tweet from Gustavo Petro in which the candidate supposedly expressed himself in a rude manner. Another user rejected the publication, shared in a mocking tone, and Robertico replied: "You can be a follower of Gustavo Petro and not lose your fucking sense of humor".
Within hours, his account, with more than 19,000 followers, was permanently suspended for repeated violations of Twitter rules. The user appealed, trying to explain that in Colombia the word "marica" is often used without the homophobic connotation that its use in other contexts may have. However, the platform stood by its decision.
Faced with Twitter's refusal, Mario Roberto contacted the company by e-mail and explained, once again, the meaning of his comment. Without getting any response, a few hours later his account was reinstated.
This is not the first time that social media have sanctioned users in Colombia for the use of that word. Last year the Oversight Board, a body that functions as a Supreme Court of Meta, had to decide on the case of a publication in which some protesters, in the framework of the national strike, told President Iván Duque to "stop being a faggot on TV". The content had been removed and different organizations, such as the Foundation for Press Freedom, Karisma and El Veinte asked the Council to revert the decision, explaining that in Colombia the connotation of the term "marica" depended on the context in which it was used and was not necessarily derogatory or discriminatory. The agency finally ordered the reinstatement of the publication, although for different reasons.