Algorithmic transparency as a fundamental right in the democratic rule of law: a comparative approach to regulation in European, North American, and Brazilian Contexts

Abstract

This article scrutinizes the escalating apprehensions surrounding algorithmic transparency, positing it as a pivotal facet for ethics and accountability in the development and deployment of artificial intelligence (AI) systems. By delving into legislative and regulatory initiatives across various jurisdictions, the article discerns how different countries and regions endeavor to institute guidelines fostering ethical and responsible AI systems. Within the United States, both the US Algorithmic Accountability Act of 2022 and The European Artificial Intelligence Act share a common objective of establishing governance frameworks to hold errant entities accountable, ensuring the ethical, legal, and secure implementation of AI systems. A key emphasis in both legislations is placed on algorithmic transparency and elucidation of system functionalities, with the overarching goal of instilling accountability in AI operations. This examination extends to Brazil, where legislative proposals such as PL 2.338/2023 grapple with the intricacies of AI deployment and algorithmic transparency. Furthermore, PEC 29/2023 endeavors to enshrine algorithmic transparency as a fundamental right, recognizing its pivotal role in safeguarding users' mental integrity in the face of advancing neurotechnology and algorithmic utilization. To ascertain the approaches adopted by Europe, the United States, and Brazil in realizing the concept of Algorithmic Transparency in AI systems employed for decision-making, a comparative and deductive methodology is employed. This methodology aligns with bibliographical analysis, incorporating legal doctrines, legislative texts, and jurisprudential considerations from the respective legal systems. The analysis encompasses Algorithmic Transparency, Digital Due Process, and Accountability as inherent legal constructs, offering a comprehensive comparative perspective. However, the mere accessibility of source codes is deemed insufficient to guarantee effective comprehension and scrutiny by end-users. Recognizing this, the imperative of explainability in elucidating how AI systems function becomes evident, enabling citizens to comprehend the rationale behind decisions made by these systems. Legislative initiatives, exemplified by Resolution No. 332/2020 of the National Council of Justice (CNJ), underscore the acknowledgment of the imperative for transparency and accountability in AI systems utilized within the Judiciary.