Constitutional Presuppositions on the Use of AI in the Executive Branch

ABSTRACT The purpose of this study is to make an approximation to the constitutionality of the use of AI systems by the Executive Branch, with special reference to the principles of action that should govern administrative activity when it is carried out with the support of these tools. The application of AI systems in the field of administrative activity must be interpreted in a positive key, due to the innumerable advantages it brings to the quality of public services in terms of effectiveness and efficiency. The different algorithmic models, with greater or lesser autonomy, are prepared to support administrative decision-making or produce automated decisions. Their role will depend on the translation of regulatory regulations into precise instructions, and on the suitability of the reference data provided to the system for training. However, the subjection of administrative activity to the principle of legality and its judicial control implies the inexcusable duty to respect all the fundamental rights involved. In the case of the use of AI systems, the protection of personal data, access to public information derived from the principle of transparency, the duty to state reasons for administrative acts derived from the principle of legality, and the principle of non-discrimination are particularly relevant. Due to these circumstances, the greatest risk presented by AI in the field of the Public Sector, and the main reason for its stigmatisation, is a consequence of the opacity of algorithmic systems and the complexity of their operation. In any case, the challenge they pose should not be faced from the exclusion of their use, but from the implementation of greater and better security mechanisms to guarantee effective protection of all the rights of the subjects involved.

KEYWORDS Artificial intelligence - Public Administration - Automated administrative decision-making - Algorithmic transparency - Fundamental rights

TABLE OF CONTENTS 1. Introduction. – 2. Constitutional Provisions Applicable to the Use of AI Systems by the Executive Branch and Principles of Action in the Field of Public Administration. – 2.1. Principle of Legality and Judicial Control. – 2.2. Principle of Transparency. – 2.3. Principle of Equal Treatment and Non-Discrimination. – 2.4. Protection of Personal Data. – 2.5. Recognition and Protection of Digital Rights. – 3. Legal Relevance of the Use of AI Systems in the Field of Public Administration. – 3.1. Conceptual Approach and Planning. – 3.2. AI Systems and Administrative Decisions. – 3.3. Automated Administrative Decisions. – 3.4. Impact of Algorithmic Models on Automated Administrative Decisions. – 3.4.1. Conceptual Approach. – 3.4.2. Legal Nature. – 3.4.3. Projection on the Duty to State Reasons for Administrative Acts. – 3.4.4. Impact on the Principles of Non-Discrimination and Transparency That Govern the Actions of the Public Administration.