From the right to know to algorithmic transparency: three days to rethink the Administration of the future with the Council and the UIMP

The meeting, which had a great success of calling with more than 60 registered students, involved specialists in transparency, technology, data protection... The panellists stressed, on the one hand, that in a democratic state it is necessary to ensure algorithmic transparency and, on the other, that the use of AI opens up huge possibilities for improving the accountability, integrity and transparency of the public sector.

The President of the Council opened the course, Concepción Campos, the Vice-Chancellor for Programming of the UIMP, Salvador Torodo, and the General Director of Administrative Simplification, Transparency and Citizen Participation of Government of Cantabria, Paulino Sánchez, which highlighted the undoubted power of AI to improve public administrations and the services they provide to the citizen, and that the challenge is to ensure that citizens continue to understand how decisions affecting their lives are made.

The President of the Council recalled in her presentation “Transparency and algorithms: legal guarantees in public decision-making”, with which the course was opened, that “in the digital world the same guarantees of citizens’ rights must be preserved as in the offline world”. We must bear in mind that the technology that serves to make automated decisions for administrations may not be neutral and biased. Therefore, he said, “we have to ensure that citizens understand how they work, that there is no arbitrariness and that administrative acts continue to be motivated. We must avoid uncritical automation.” He advocated extending the obligations of active publicity and transparency from design, as well as harnessing the potential of technology to combat the abuse of administrative silence in the face of requests for public information. “From the point of view of transparency, we have gained from technology”, he defended, and recalled that the Council aims to be a close reference and an aid for the Administrations to comply with their legal obligations in this matter.

In his presentation, the chairperson de la Spanish Data Protection Agency (AEPD), Lorenzo Cotino, defended the need for the supervisory authorities in the field of personal data protection, transparency and AI (AEPD, CTBG and AESIA, respectively) to develop common criteria that guide the Public Administrations to guarantee the necessary data protection, traceability and explicability of decisions based on artificial intelligence that, he recalled, will be as essential “as breathing”, since it exponentially increases the capacities and efficiency of organizations.

For his part, the Doctor Engineer of Telecommunication Manuel J. García Rodríguez (Banco de España) stressed that the AAPP must design and tender their computer systems incorporating the principle of transparency and the obligation of explicability in the documents themselves. He also opted to clarify the competencies and promote coordination between the different levels of administration and the guarantee bodies, as well as for the creation of a constantly updated, dynamic source code repository that allows to compare the versions of administrative applications that use automation or AI, as algorithm records.

“Document management is a critical infrastructure for transparency.” With this statement the debate was opened in the session that closed the first day, with the participation of the Deputy Deputy Director of Transparency and Good Governance of CTBG, Julian Prior, the technical adviser to the General Subdirectorate of State Claims, Carmen Rial and the Deputy Assistant Director of State Level Claims, José Santiago Fernández. The Administrations, they stressed, produce information and documentation at an increasingly dizzying pace and the correct use of AI is key to make all this information available to the city and avoid opacity, “since producing information is not the same as communicating it and transmitting it efficiently to the citizen.” In the practical session, they detailed the risks and opportunities that AI poses throughout the document management cycle, insisting on the importance of alliance and collaboration between document managers, ICT staff and legal, to ensure the conservation of information, now in digital format, in the future.

The second day of the meeting was opened Deputy Director General for Transparency and Good Governance, Gonzalo Gómez de Villalobos, which addressed how AI can serve to enhance one of the two pillars of transparency in Spain: active advertising. He detailed how technology can be harnessed to improve the information published by PPPs for accountability and also to facilitate the guarantee bodies’ own work in assessing compliance with these obligations. He announced that the Council is working on updating its MESTA evaluation methodology and on the development of a guide to help obliged subjects on active advertising, among other projects.

How AI is affecting the second pillar that makes up the transparency system in Spain, the right of access to public information, focused the intervention of the Deputy Director General of State Level Claims of the CTBG, Marta Timón. “We are entering a more complex area than the one we had,” he said, “where there are more and more requests for access to information related to algorithms, models and data,” which will require a “very fine weighting” effort. He also talked about the increase and changes that AI can bring both in the requests for information received by the Administrations and in the claims received by the guarantee bodies. This entails risks, he said, but also the opportunity to use technology to improve how the Administration satisfies the exercise of the right, without forgetting the need to maintain traditional channels for vulnerable groups affected by the digital divide.

For its part, the Deputy Director General of Claims of Autonomous Communities and Local Entities of the CTBG, José Manuel Jover, He agreed that AI is a tool that can be used by administrations to improve the response to requests for information that, currently, in a very high percentage, are rejected by administrative silence (45% of the state-level claims resolved by the Council in 2025 had been filed for this cause; percentage that rises to 75% in the autonomous and local area of those CC.AA. which have conferred on the Council the exercise of this competence). In her presentation she gave a practical demonstration of how AI can, in those simpler claims, be of great help through automation and its predictive use.

Transparency as a means of guarantee for an ethical and responsible use of AI was the guiding thread of the presentation of the professor of Philosophy of Law and dean of the Faculty of Law of the University of Seville, Fernando Llano. “When you decide a machine, transparency is no longer just about showing the record: it is about the citizen understanding, monitoring and reproducing the logic of the decision, and also being able to challenge it.” Algorithmic opacity turns the citizen into a number; no automated act is motivated if its logic is not verifiable; opacity generates arbitrariness and that is why transparency is so important”, to ensure the rights of citizens, he said.

For its part, María Esteller, Assistant Director General for Reports and Test Infrastructures of the State Agency for Artificial Intelligence Supervision (AESIA), detailed the work and functions of this body to supervise and promote the culture of artificial intelligence in Spain. Esteller also broke down the transparency obligations set out in Article 50 of the European AI Regulation, which enters into force in August, explaining the importance of the explainability of decisions supported or based on technological tools.

The vision of civil society on transparency and AI came from several organizations and professionals very active in defending the right to know, who presented different initiatives that they have developed in this area. Intervened Luis Cayo Pérez, president of Spanish Committee of Representatives of Persons with Disabilities (CERMI); David Cabo, co-founder of the Fundación Civio; David Fernández, computer engineer, creator of the Open Contracting and Info platforms; Germán J. Arenas, researcher at the Hay Derecho Foundation, and Mónica Roa, researcher at Transparency International Spain. They stressed the need for the Administration to incorporate personnel with technological knowledge, to improve accessibility to information from the “right to understand” approach, and to enable an inventory that collects applications and systems that incorporate automated decisions. They also called for the strengthening of the resources both of those administrative units that are responsible for transparency and of the guarantor bodies themselves, which, they stressed, must be endowed with sanctioning powers.

The last day of the course was opened with a discussion table on transparency and AI at European level, in which Community regulation and governance in this matter were compared with other models. “The European system is a guarantor and an anthropocentric system,” summarized the General Secretary of the CTBG, Isabel Lombardero. For his part, the president of Public Sector and Artificial Intelligence Observatory, Gabriele Vestri, warned that “AI in certain qualities has already surpassed us humans” and opted for the need for public bodies to provide codes of good use, “which do not exist in most cases” and ethics committees in the matter, since AI “is affecting our fundamental rights”. María Pascual, specialist in Public Policy, Anti-Corruption, Integrity and Open Government of the OECD, spoke of the AI principles adopted by this multilateral body as a guide to the interpretation of the European AI Regulation, which recognizes the right “to obtain a clear and meaningful explanation when an automated decision impacts our lives.”

The last session of the course addressed the new legal scenarios that open up to AI. The Director of Cabinet of CTBG, Isabel Jiménez, analyzed the challenges that technology and, in particular, artificial intelligence, display over the traditional legal-administrative categories on which our State model is based and the need to examine them for their preservation. For its part, the Judge of the Supreme Court Juan Pedro Quintana He explained in detail the relevant BOSCO judgment, for which he was rapporteur, and stressed the importance of ensuring the explicability of automated decisions, since “it is difficult to combat the biases that algorithms may have with opacity”, without necessarily requiring, in all cases, access to the source code, which will require weighing. Along with this, he explained the value of the concept of “humanity reserve.” La Professor of Administrative Law at the University of Castilla-La Mancha, Susana de la Sierra, he stressed that AI is a transversal technology, and that, therefore, more than a sectoral regulation, “a transversal reflection to see how the necessary legal and regulatory guarantees are introduced”. Along with this, he stressed that today the control of the Public Administration is dual, which calls for the creation of instruments of cooperation between all the agencies that carry out this function; he also advocated the empowerment of specific mechanisms so that citizens can resort to the acts in which the algorithms have intervened.

La scivil service secretary, Consuelo Sánchez, closed the meeting, stressing that “transparency is no longer about publishing documents, it is about explaining complex processes” and one of the challenges is “to move from formal transparency to real transparency”. “AI is not the end of transparency, it is the greatest test of its maturity,” he said, and said that “AI can transform the Administration, but only transparency can ensure that this transformation does not imply loss of democratic control.”