Back to AI: Smart law or artificial intelligence for law?

Abstract

This essay examines the legal challenges posed by the incorporation of artificial intelligence (AI) into the Guatemalan legal system, with a particular emphasis on the administration of justice. It is based on the hypothesis that the adoption of AI will only be constitutionally legitimate if implemented within a framework of safeguards, adapted to the country's institutional reality. The research employs a qualitative legal-analytical and comparative approach, which includes the study of international regulatory frameworks—such as the European Regulation on Artificial Intelligence (EU Regulation 2024/1689)—and the analysis of relevant case law, such as Judgment T-323/2024 of the Colombian Constitutional Court (use of ChatGPT in judicial decisions), the ruling in Denegri v. Google of the Argentine Supreme Court (right to digital oblivion), Appeal 27/2021 of the Mexican Supreme Court (Pegasus case), and File 3552-2014 of the Guatemalan Constitutional Court, which incipiently recognizes informational self-determination. The results identify four main regulatory gaps: lags in personal data protection, absence of public policies and ethical frameworks on AI, institutional weaknesses in cybersecurity, and an insufficient framework for cybercrime. The proposal is to create a Law on Artificial Intelligence and Protection of Rights, based on principles of transparency, explainability, and accountability, along with a National AI Strategy for Justice that includes specialized training, controlled pilot projects, and democratic oversight mechanisms. It is concluded that modernizing the Guatemalan justice system requires an approach that balances technological innovation and effective protection of fundamental rights.