International

The director of the RAE asks to take care of Spanish in AI so that they do not create their own dialects

The director of the Royal Spanish Academy (RAE) and president of the Association of Academies of the Spanish Language (Asale), Santiago Muñoz Machado, called for taking care of the development of artificial intelligence (AI) so that machines speak Spanish in the most correct way possible and at the same time prevent their algorithms from creating digital dialects that would constitute “a great break for the unity of the language”.

“It is important that machines, which are a growing number of non-human individuals who use our language, do so in the best possible way,” Muñoz Machado said in an interview with EFE during the celebration in Quito of the XVII Congress of the Asale, in which issues such as the challenges and opportunities of Spanish in the face of new technologies and the digital world will be discussed.

The Spanish jurist considered that for this there is “a competitive advantage: none of the technology companies that have machines that speak is interested in speaking badly, in the same way that no school or university is interested in having students who do not handle the language well.”

“Large technology companies are not interested in having illiterate people. To help them not have them, what we tell them is that, when they teach the machines to speak Spanish, they use the tools of the language academies, and in that way the machines will speak a language just like that of humans,” explained Muñoz Machado.

Spanish and AI

The director of the RAE admitted that, for the moment, they have not detected major deviations in artificial intelligence, with respect to the canon that the academies establish.

20260224_estafa_mh_728x90

“(But) we must take care of this and that it remains so, because it would be feasible for the algorithms that handle artificial intelligence to determine variations of the language that would ultimately create digital dialects of artificial intelligence, not unintelligible from the common Spanish language,” warned Muñoz Machado.

“That would be a great breakdown of the unity of the language and an unbearable injury to a language that they speak now and with which 600 million people understand. It would be very serious if it happened, but I think there is no will or interest in it happening, not even economically,” he concluded.

In that sense, he emphasized that the scenario that these machines could create languages derived from Spanish, as well as other languages, that only they understood is not science fiction, because “everything that comes out in this matter is happening now.”

AI to detect new words

The jurist, who has directed the RAE since 2018, recalled that in the previous Asale congress, held in 2019 in Seville (Spain), the Spanish Language and Artificial Intelligence (LEIA) project was already created.

“In recent years we have had great revolutions that can affect the tools we use for the general regulation of the language, especially the digital revolution and artificial intelligence. We have opened up to them immediately,” he said.

20260224_estafa_mh_728x90

Thus, they have also seen with AI an opportunity to use this technology to detect new terms and forms of expression that arise and that do not reach them through traditional channels.

The digital revolution has also led to the appearance of new words to name elements or phenomena that did not exist before, most Anglo-Saxon terms that do not generate fear in the academies of the Spanish language.

A mixed language

“It is not a great tragedy because Spanish has always been mestizo, a language very given to incorporating terms from other languages from its origins. He already incorporated many Arabic words and other neighboring languages. And in countries like the Americans it has terms of their original languages,” Muñoz Machado said.

“We’re not too worried. The ‘Dictionary of the Spanish Language’ has 94,000 entries and 189,000 meanings. Every year we incorporate a maximum of a dozen new terms from English, and we also incorporate them raw, without Castilianizing or Spanishizing them, in the same way they mean in English. We enrich the language in that way and nothing happens,” he added.

However, the director of the RAE emphasized that “a different thing is to use English expressions that have equivalent in Spanish only for snobbery or to show a certain knowledge of that culture in an unnecessary and harmful way for the quality and integrity of our language”

20260224_estafa_mh_728x90

At the Asale congress, among other novelties, a ‘Pan-Hispanic Guide to Clear Language’ will also be presented, with the aim that the official communications of the institutions can be understood by any citizen and that they can thus exercise “the right to understand,” as Muñoz Machado defined it, “which is the basis of the exercise of other rights.”

International

Brazil offers to mediate Colombia-Ecuador tensions, calls for restraint

The government of Brazil has offered to mediate in the ongoing tensions between Colombia and Ecuador, while calling on both nations to exercise restraint.

In a statement released Wednesday, Brazil’s Ministry of Foreign Affairs urged the parties involved to act with moderation and seek a peaceful resolution to the dispute.

“Brazil encourages all sides to act with moderation in order to find a peaceful solution to the controversy. It stands ready to support dialogue efforts aimed at preserving peace and security in the region,” the statement said.

Brazil also expressed “serious concern” over reports of deaths in the border area between Colombia and Ecuador, noting that the circumstances surrounding the incidents have not yet been clarified.

The diplomatic move comes amid rising tensions between the neighboring countries, increasing regional concern over stability and security along their shared border.

20260224_estafa_mh_728x90

International

U.S. lowers travel advisory for much of Venezuela but keeps high-risk zones under warning

The U.S. Department of State announced on Thursday that it has lowered its travel advisory for much of Venezuela to Level 3 (“Reconsider Travel”), reflecting what it described as improved security conditions in parts of the country.

However, the agency will maintain the highest Level 4 warning (“Do Not Travel”) for several regions, including the states of Táchira, Amazonas, Apure, Aragua and Guárico, as well as rural areas of Bolívar, citing ongoing risks such as crime, kidnapping and terrorism.

The updated advisory marks a shift from December, when the United States raised the alert for Venezuela to Level 4 nationwide, warning of severe security threats.

Despite the partial downgrade, U.S. authorities continue to urge caution, emphasizing that conditions remain volatile in certain areas and that travelers should carefully assess risks before planning any trips to the country.

International

EU lawmakers move to ban AI tools that generate non-consensual nude images

Members of the European Parliament are pushing to ban across the bloc artificial intelligence services that allow users to digitally “undress” people without their consent.

The proposal, adopted on Wednesday at committee level, aims to prohibit applications that generate non-consensual explicit images. Irish lawmaker Michael McNamara, one of the sponsors, said the measure seeks to stop tools that “have caused significant harm for the benefit of a few.”

Dutch MEP Kim van Sparrentak welcomed the move, calling it “a major victory, especially for women and children in Europe.”

The amendment, part of broader EU legislation on artificial intelligence, was approved by the Parliament’s civil liberties and internal market committees. It specifically targets systems that use AI to create or manipulate sexually explicit or intimate images resembling identifiable individuals without their consent.

The proposal will be put to a full vote in the European Parliament on March 26. If adopted, lawmakers and European Union member states will need to agree on a final version before it can take effect.

20260224_estafa_mh_728x90

Separately, representatives of the 27 EU countries recently backed a Franco-Spanish amendment seeking to ban AI services used to generate non-consensual sexual images or child sexual abuse material.

The initiative follows controversy surrounding a feature introduced in Grok, developed by xAI, which allowed users to create simulated nude images from real photos. The tool sparked widespread criticism and prompted an EU investigation.

In response, xAI restricted image generation features in mid-January to paying subscribers and stated it blocks the creation of sexualized images in jurisdictions where such content is illegal.