The RAE Advances on the Shoulders of the Giant of Artificial Intelligence

In 2019, the Royal Spanish Academy (RAE) climbed onto the shoulders of the giant of artificial intelligence. The three-century-old institution knows it must remain there to drive and ensure the future of Spanish in the rapid era of algorithms, platforms, and applications to digitally fix, clean, and enhance our language.

To achieve this, it modifies and enhances the digital tools of the LEIA system (Spanish Language and Artificial Intelligence), a happy acronym that alludes to the galactic princess. Among them, a Words Observatory capable of 'filtering' a million expressions a day, and linguistic verification systems to detect spelling, lexical, and grammatical errors and compare, playfully, the variations of Spanish worldwide.

"Artificial intelligence is not an option or a whim. It is a necessity; an indispensable ally to preserve, study, defend, and project the Spanish language into the future in the digital universe," states the director of the learned house, Santiago Muñoz Machado.

He has piloted a titanic digital effort in which nearly five million euros have been invested, "more than necessary" for the future of the language spoken today by 600 million people. Soon, its tools will be available to all Spanish speakers through the RAE's website.

"We will continue to stand on the shoulders of that AI giant to advance from there, which is now a commitment of the RAE," said Aleida Alcaide, Director General of Artificial Intelligence at the Ministry for Digital Transformation, whose contribution has been crucial. Aleida presented the five renewed LEIA tools that will facilitate this path in the digital age alongside the expert and project director, Asunción Gómez-Pérez.

Hosted on a private cloud, the LEIA platform integrates a Spanish Words Observatory that detects neologisms, derivatives, technical terms, regionalisms, and foreign words. AI provides the capacity to analyze a million forms and expressions daily in the press, social networks, blogs, bulletins, and many other sources. It detects words not registered in the Spanish Language Dictionary (DLE) – such as 'random', 'juernes', 'youtuber', or 'podcast', for example – whose use may raise doubts among Spanish speakers.

Another part of the renewed LEIA project is a linguistic query system that collects the doubts and questions that reach the RAE through its website and the social network X. AI automatically classifies them according to their lexical, morphological, or syntactic category, reducing the response times of RAE linguists. If the system cannot provide a conclusive answer, the user will receive a "personalized" explanation from RAE linguists, so it is not a "chatbot," the RAE highlights.

An automatic linguistic verification system includes the detection and classification of spelling, lexical, syntactic, and morphological errors that are analyzed when entering a text. Without ever making automatic corrections, it offers correction suggestions and explanations through links to academic doctrine.

The Spanish Linguistic Diversity Compiler compares linguistic varieties from different geographical areas of Spanish. It allows massive capture and analysis of linguistic information from Spanish speakers through games or challenges.

To trace this radiography of the rich diversity of our language, an interactive section has been created for speakers to report on their use of the language. Through an intuitive interface, the system poses questions, called challenges, for which it presents a word, a text, an image, or an audio. "Where did you learn to speak Spanish?" asks this system that the RAE has tested with 'datathons' (data collection marathons) and with fifteen hundred fifth-grade children.

Also, a digital archive has been created by digitizing the so-called 'red dot cards', lexicographic slips of "great value" that designate the first documentation available for each word.

The powerful and renewed tools are already used in their daily work by the RAE teams and "soon" will be available for free and open use by the entire Spanish-speaking community through the RAE's website, according to the institution without committing to a date.

"The RAE has been one of the first cultural institutions in the world to recognize the need to integrate artificial intelligence into linguistic work, and that is why we have a privileged position. We cannot conceive our work without it. It is an indispensable ally to preserve and project the Spanish language," reiterated Muñoz Machado, who managed to involve major technology companies such as Google, Microsoft, Telefónica, or AWS in the endeavor.

"Our language must be alive not only in books but also in algorithms, platforms, and applications," affirmed Aleida Alcaide.