IA y revisión judicial: nuevos desafíos para el control de decisiones automatizadas

IA y revisión judicial: nuevos desafíos para el control de decisiones automatizadas

AI and judicial review: new challenges for the control of automated decisions

by Clara Carolina Romero Borja · View the channel on PSI

Humanities, Social Sciences & Thought Law & Justice LNQDigital, IT and Communications lawLNQHArtificial intelligence law and regulations
🎙 Clara Carolina Romero Borja 👥 2K 📅 September 26, 2026 ⏱ 53 min 👁 9 📄 expert opinion 🧭 2026-09-27
Available in: English (current) Français

Keywords

automated decision-makingAI and lawjudicial controladministrative procedurealgorithmic accountability

Summary

The talk, delivered by Dr. Clara Carolina Romero Borja, examines the challenges posed by artificial intelligence to judicial review of administrative decisions. She begins with a foundational question: can a judge control an administrative decision if the process of how it was made cannot be reconstructed? She traces the evolution from traditional and electronic administration to algorithmic and fully automated administration, arguing that the legal implications differ significantly across these stages. She distinguishes four levels of AI intervention: as a tool, as assistance, as a determining element, and as fully automated decision-making, noting that judicial review cannot be uniform across these levels. She highlights two recent U.S. cases (Douglas v. Deutsche Bank and a Second Circuit case) where lawyers relied on AI-generated hallucinations, leading to sanctions and emphasizing that AI cannot displace human responsibility. She introduces the concept of a ‘black box’ — when an automated system denies a citizen’s request without offering explainable reasons, making judicial control structurally difficult. The speaker calls for a nuanced legal framework that distinguishes between human validation and genuine human decision-making, and stresses the need for public administration to ensure accountability and due process in the age of AI.

195 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value lies in the speaker’s practical experience as a judge and her clear conceptual framework, which is useful for legal scholars and practitioners. The argumentation is well-structured: she moves from concrete cases to theoretical stages, and argues strongly that the shift to automated decision-making undermines traditional guarantees of administrative law. She presents a compelling analogy between lawyers’ blind trust in AI and the risks for public officials, and she distinguishes levels of intervention to avoid a blanket rejection of AI. The reasoning is sound, though it relies on anecdotal rather than systematic evidence, and the discussion of solutions (e.g., how to ensure transparency or auditability) remains at a general level.

Scientific Rigor, Source Quality, Title Accuracy

The speaker references the UNESCO guidance on AI and justice and mentions at least two specific U.S. court cases, but provides no direct citations or links; this restricts verifiability. The title is well-aligned with the content, as the talk consistently addresses judicial review of automated decisions. The presentation’s rigor is moderate: the legal reasoning is current and informed, but the absence of documented sources and the informal delivery lower its scholarly weight. No audience comments were provided, so no public feedback analysis is possible.

210 words

Title / Content Match

The title accurately reflects the content, which focuses on how AI challenges judicial review of automated administrative decisions.

Quality & Reliability

7/10

The speaker is a judge in the Administrative Contentious Chamber of the Ecuadorian National Court, providing substantive legal expertise. The presentation draws on specific real cases and international frameworks (UNESCO), with a clear conceptual framework. However, the talk is conversational and lacks written sources or citations for several claims, and the low production values and audience limit peer review.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • Douglas v. Deutsche Bank National Trust — U.S. District of Columbia Court of Appeal case where AI-generated fictitious caselaw was submitted, leading to sanctions.
  • Second Circuit case (including lawyer Elliot Shields) — A case involving a NYC Metro accident, where the lawyer confessed to not reviewing AI-generated arguments.
  • UNESCO Guidance on AI and Justice — Referenced as background for global AI use in judicial systems.

Contribution & Novelties

The talk provides a nuanced, practice-oriented analysis of how automated administrative decisions complicate judicial review, proposing a staged distinction between algorithmic assistance and full automation. Its major contribution is highlighting the ‘black box’ problem in administrative contexts and the subtle shift from genuine human decision to mere validation of AI output.

Pour aller plus loin :

  • Automated decision-making — Overview of how automated systems affect administrative processes and the stakes for legal accountability.
  • Algorithmic bias — Explores how biased algorithms can lead to discriminatory administrative decisions, a core background to the speaker’s concern.
  • Judicial review — Provides the general legal concept within which the talk situates the challenges.
  • Right to explanation — A key principle in AI regulation (especially GDPR) that directly relates to the need to understand automated decisions for judicial control.

133 words

Radar Profile

The radar profile shows moderate-to-high values across all four dimensions, with quantity and quality of information scoring highest, while the overall reliability is slightly lower due to the lack of verifiable references. The technicality level is solid for a legal audience, but not excessively formal.

Reliability 7/10