Support Centre

You have out of 5 free articles left for the month

Signup for a trial to access unlimited content.

Start Trial

Continue reading on DataGuidance with:

Free Member

Limited Articles

Create an account to continue accessing select articles, resources, and guidance notes.

Free Trial

Unlimited Access

Start your free trial to access unlimited articles, resources, guidance notes, and workspaces.

Spain: Agency for the supervision of AI - overview

Miguel Valdés Borruey, Luis M. Latasa Vassallo, and María Ñíguez Olalla, from EJASO, provide a concise overview of Spain's pioneering efforts in artificial intelligence (AI). They delve into the creation of the Artificial Intelligence Supervisory Agency (AESIA), its organizational structure, and Spain's commitment to ethical AI practices within the broader EU context.

Nadya So / Essentials collection / istockphoto.com

Spain has become the first EU Member State to approve the creation of AESIA in anticipation of the future Artificial Intelligence Act (AI Act). AESIA is scheduled to commence its supervisory activity no later than December 3, 2023.

The proposal for the AI Act was published by the European Commission on March 21, 2021, and it is currently in the negotiation process. In accordance with Article 59 of the proposal, each Member State will be required to establish or designate a national authority with the authority to ensure the proper application and enforcement of the future Regulation. According to the definition provided in the text, a 'national supervisory authority' refers to 'the authority to which a Member State assigns the responsibility for the implementation and application of [the future AI] Regulation, for coordinating the activities entrusted to that Member State, for acting as the single contact point for the Commission, and for representing the Member State at the European Artificial Intelligence Board.'

In this context, Spain allocated funds for the AESIA in the 2022 National Budget, anticipating the approval of the AI Regulation. Meanwhile, the seventh additional provision of the Spanish Start-ups Act (Law 28/2022 of 21 December on action to foster the emerging enterprises ecosystem) endorsed the establishment of AESIA. Finally, the Statute of the Spanish Artificial Intelligence Supervisory Agency (AESIA Statute) was approved by Royal Decree 729/2023 on August 22, 2023. Its operational commencement will coincide with the formation of the Governing Board and the appointment of the Director.

AESIA is affiliated with the Spanish Ministry of Economic Affairs and Digital Transformation through the Office of the Secretary of State for Digitalization and Artificial Intelligence. The Agency will be located in the city of La Coruña. In this capacity, AESIA has assumed the role of the Spanish national authority tasked with overseeing and ensuring appropriate implementation of all European and domestic legislation governing the development and use of AI systems and, specifically, of algorithms.

AESIA's purpose is, therefore, to prepare the Spanish administration to fulfill the obligations and responsibilities established by the future AI Regulation. Additionally, it aims to guide the Spanish AI ecosystem until European legislation is established, encouraging voluntary supervision through quality seals. Moreover, AESIA will create AI regulatory sandboxes, offering AI innovators and systems developers a safe space in which to work.[1]

Spain's new AESIA is an autonomous public body with its own legal personality and financial resources. Its key functions include inspection, verification, and the imposition of sanctions in matters related to the secure, responsible use of AI systems. However, AESIA's duties will not impinge upon the remit of other government agencies responsible for health and labor relations matters.

AESIA's actions will be governed by the purposes outlined in Article 4 of the AESIA Statute, which include:

  • raising awareness, disseminating information, providing training, and promoting the responsible, sustainable, and accountable development and use of AI;
  • defining advisory and alert mechanisms for society as a whole and the entities involved in the development and use of AI;
  • cooperating and coordinating with other national and supranational AI supervisory authorities;
  • establishing regulatory sandboxes to reinforce the protection of users and prevent discriminatory bias; and
  • overseeing the start-up, use, and marketing of systems utilizing AI, especially where they may pose significant risks to health, safety, and fundamental rights.

In accordance with Article 8 of the AESIA Statute, the Agency will be governed by the general interest and obligated to adhere to certain fundamental principles in all its activities, including, among others, autonomy of action, technical independence, transparency and good governance, accountability, effectiveness, efficient use of resources and continuous assessment of the quality of its management processes and procedures, institutional cooperation, and equal treatment of men and women. These basic principles also include institutional cooperation, which involves the search for synergies in cooperation with other official Spanish and international entities, agents, and institutions, whether public or private, to promote the spread of knowledge in all areas.

Notwithstanding any other functions that may be entrusted to AESIA under future EU and Spanish legislation, its competencies can be summarized as follows:

Support innovation via regulatory sandboxes

Promotion of test environments to ensure proper alignment of innovative AI systems with prevailing legislative frameworks.

Foster the use of sustainable algorithms

Support for the development and use of AI systems from a dual environmental and sustainable perspective.

Support an AI gender perspective without discriminatory bias of any kind

Fostering the development and use of AI systems embodying the principle of equal opportunity between women and men in their design and execution, and promotion of impact assessments capable of detecting possible cases of discriminatory bias of any kind prohibited by law.

Support efforts to build confidence in AI

Creation of a voluntary certification framework allowing private entities to guarantee responsible design of digital solutions and provide assurance with regard to their technical solutions, thereby avoiding overregulation and leaving room for innovation.

Raise awareness of ethical and humanist AI

Fostering the dissemination of knowledge and training in relation to ethical, humanist AI to reveal the potential, challenges, and uncertainties arising with regard to the development and use of AI.

Public-private partnership

Support for public-private partnership to encourage the creation of cooperative frameworks in the AI industry in order to encourage humanist development and proper use of the technology by industry and business.

Regulatory compliance

Supervision and sanctions, where appropriate, are in line with European legislation governing oversight of AI systems.

Other

Instrumenting mechanisms to identify trends and assess the social impact of AI.

Helping the implementation of AI-related programs via contracts, agreements, and other legal instruments available to support programs in the field of artificial intelligence.

 

Chapter three of AESIA's Statute establishes the organizational structure of AESIA, aiming to achieve its goals and fulfill its competencies and obligations. On this basis, AESIA will be structured with governance and executive organs. The former, governance organs, will consist of the Agency's Chair and Governing Board, while the latter, executive organs, will include the Directorate, the Office of the Secretary-General, the Subdirectorate of Reports and Test Infrastructure, and the Subdirectorate of Certification, Trend Evaluation, Coordination, and AI Training.

Finally, it is important to contrast AESIA with the European Centre for Algorithmic Transparency (ECAT), another recently created entity based in Sevilla, Spain. While both authorities are destined to become the pillars of supervision and promoters of transparent, secure practices in the field of AI and algorithms, they will act in different contexts taking different approaches. Unlike AESIA, ECAT's role is to oversee compliance with the Digital Services Act (DSA), and it operates under the European Commission. The ECAT's purpose is to provide support to the Commission in its supervisory tasks and to ensure that very large platforms and search engines use transparent, reliable algorithmic systems. To guarantee a secure, predictable, and trust-based online environment, the Commission will carry out functions such as analyzing risk self-assessments and transparency reports, assessing potential systemic risks with the technical and scientific support of ECAT, and proposing new methodological approaches and best practices for algorithmic transparency to mitigate identified risks.

In conclusion, the creation of AESIA is an important milestone for the EU as a whole, emphasizing Spain's readiness and commitment to comply with future AI regulations. This commitment is aimed at ensuring an ethical and humanist approach to AI systems and innovation in alignment with fundamental rights and European values.

Miguel Valdés Borruey Partner
[email protected]
Luis M. Latasa Vassallo Partner
[email protected]
María Ñíguez Olalla Senior Associate
[email protected]
EJASO, Madrid


[1] The Spanish Government is currently working on a Royal Decree to create a regulatory environment for controlled trials (Sandbox) designed to test compliance with the draft AI Regulation.

Feedback