Los Algoritmos en los Sistemas de Administración de Justicia. Debilidades operativas. (IA)

Value of the Information & Strength of the Argument

The presentation offers valuable insights into the practical challenges of implementing AI in judicial systems, particularly in Latin America. The speaker draws on his professional experience and provides concrete examples of algorithmic bias and operational failures. However, the argumentation is largely anecdotal and lacks rigorous empirical evidence. The speaker makes broad claims without providing specific data or citations, which weakens the overall persuasiveness. The discussion of the ‘black box’ problem and the need for transparency is well-articulated, but the proposed solutions are general and lack detailed implementation strategies.

Scientific Rigor, Source Quality, Title Accuracy

The presentation demonstrates moderate scientific rigor. The speaker references well-known cases like COMPAS and Prometea, but does not provide specific sources or citations. The statistics mentioned (e.g., ‘over 15 countries’, ‘60% without auditability’) are not substantiated. The title accurately reflects the content, focusing on operational weaknesses. However, the lack of verifiable sources and the reliance on personal opinion reduce the overall reliability. The speaker’s expertise in cybersecurity lends some credibility, but the presentation would benefit from more rigorous referencing.

The title accurately reflects the content, focusing on operational weaknesses of algorithms in justice systems.

6/10

The speaker provides a general overview of algorithmic justice in Latin America, citing well-known examples like COMPAS and Prometea, but lacks detailed citations and rigorous data. The presentation is more of an expert opinion than a systematic review, with some unverified statistics and a reliance on anecdotal evidence.

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

• 02:00Introduction and speaker's background
• 05:00Definition of algorithmic justice and its promises
• 10:00Discussion of algorithmic bias and the COMPAS case
• 15:00Explanation of the 'black box' problem
• 20:00Data deficiencies and fragmentation in judicial systems
• 25:00Regulatory gaps in Latin America
• 30:00Examples of AI systems in Brazil, Argentina, and Mexico
• 35:00Institutional challenges and recommendations

• COMPAS (Correctional Offender Management Profiling for Alternative Sanctions) — Mentioned as an example of algorithmic bias in the US judicial system.
• Prometea IA (Argentina) — Cited as an AI tool used in Buenos Aires for predictive judicial decisions.
• Predpol (Mexico) — Mentioned as a predictive policing software used in Mexico.
• SEEU (Sistema Eletrônico de Execução Unificado, Brazil) — Cited as a Brazilian system for managing criminal execution processes.

• COMPAS Recidivism Risk Assessment — ProPublica investigation on COMPAS bias, supporting the speaker's claims.
• Prometea AI in Argentina — UNESCO article on Prometea, providing context on its use and limitations.

• COMPAS Recidivism Risk Assessment — Northpointe (now Equivant) disputes ProPublica's findings, arguing that COMPAS is not biased when accounting for different metrics.