LIMITES E POSSIBILIDADES DA UTILIZAÇÃO DA INTELIGÊNCIA ARTIFICIAL NO PLANEJAMENTO DAS CONTRATAÇÕES PÚBLICAS

Abstract

This study analyzes the limits and possibilities of using artificial intelligence (AI) in the planning stage of public procurement, with emphasis on Federal Law No. 14.133/2021 and the constitutional principles of Public Administration. The research is justified by the scarcity of systematic studies on the application of algorithms in this stage, although it is decisive for the efficiency and legitimacy of government purchases. The hypothesis is that AI can optimize administrative processes, reducing errors and increasing efficiency, provided that legal, technical and ethical safeguards are ensured. The methodology is qualitative, exploratory and descriptive, based on bibliographic review, documentary analysis and case law from the Federal Court of Accounts (TCU), as well as guidelines from international organizations such as the OECD and UNESCO. The analysis focused on the use of AI in preparatory documents — Demand Formalization Document, Preliminary Technical Study, Risk Map and Terms of Reference — highlighting benefits such as standardization and time savings, but also risks related to information security, explainability and respect for fundamental rights. The results indicate that AI must be regarded as a supporting tool, not as a substitute for administrative discretion. Governance guidelines were proposed, emphasizing human supervision, responsible data management, audit trails and correction mechanisms. The conclusion is that, although AI represents a concrete opportunity for modernization, its use in procurement planning is only legitimate when integrated with legality, transparency and the protection of fundamental rights. Keywords: Public Administration. Artificial Intelligence. Planning. Public Procurement. Algorithmic Governance.