In December 2021 we started talking about the Spanish Artificial Intelligence Supervision Agency (AESIA). The organization was created following a proposal that sought to create “a State Agency for Algorithms that regulate, for example, their use in their application to work, study their impact on mental health and determine if they can be discriminatory.” It seems that the agency is preparing to finally begin operations in 2025. The question remains the same: what will it actually do.
Headquarters in La Coruña. The evolution of the project has been slow. In September 2022 we learned that the AESIA headquarters would not be in Madrid, and after a selection process, La Coruña was chosen as the city where its facilities would be located. The agency will occupy the emblematic La Terraza building.
80 employees in 2025. As indicated in Expansión, AESIA plans to hire up to 80 employees between this year and next, and is already carrying out the necessary selection processes for this. The Government has reported in a written response to the Congress of Deputies that the agency already has a general director, a general secretary, two division heads, a deputy director and administration staff.
Theoretical start of the activity. In this economic newspaper they also indicate that AESIA will have the capacity to inspect prohibited practices in AI starting next February 2, 2025.
Source: La Moncloa.
Will supervise and sanction. According to that Government communication cited in Expansión, AESIA’s role will be regulatory and supervisory, but also sanctioning. In fact, as of August 2, “it will assume full sanctioning power and other governance powers established by community regulations.” Its roadmap, published as a PDF on La Moncloa’s website, arrives in mid-2027 and defines objectives as a code of good practices for AI models.
What legislation will apply? In that presentation from La Moncloa it is indicated that the agency will be responsible for the “development of the functions assigned in the European regulation”, which means that it will indeed focus on the application of the EU AI Law. Specifically, it will “monitor compliance with the requirements of high-risk AI systems or the transparency requirements of minimum-risk systems.”
One of the slides from the recent AESIA presentation. Source: La Moncloa.
The Spanish strategy in AI. The Artificial Intelligence Strategy 2024, approved in May, proposed an investment of 1.5 billion euros to facilitate the adoption of AI in the public and private sectors. Among its objectives is the promotion of a “responsible and humanistic” AI that would precisely be articulated by the AESIA.
What will this organization do? According to the BOE, the AESIA will carry out “measures aimed at minimizing significant risks to the safety and health of people, as well as their fundamental rights, that may arise from the use of AI systems.” The organization will be attached to the Secretary of State for Digitalization and Artificial Intelligence (SEDIA). In a press release from June 2024, the Congress of Deputies added that the AESIA will guarantee “the alignment of the algorithms with the requirements of the Community Regulation of Digital Services.”
We don’t really know what all this means.. Spain is not particularly leading in AI, although there are success stories like Freepik, but here we find an agency that will theoretically become a supervisory and regulatory body for AI. The question, of course, is what AI will regulate. Probably not so much the one developed in the EU – because there are not as many platforms – as the one that comes from countries like the US or China, where development and innovation are frenetic.
Back to regulatory obsession. The orientation seems to be very much in line with the regulatory obsession of the European Union, which is causing two types of AI to gradually exist: one layer for Europeans, and the other, with all its capacity, available to the rest of the world. world. Apple Intelligence, for example, has delayed its arrival in the EU for this reason, and the same is true for many other AI platforms.
Image | Wikimedia Foundation
In WorldOfSoftware | Apple is playing an important game of chess with the EU and its regulation. And for now he is winning
