Whole Of Case Visibility
Whole of Case Visibility refers to the requirement for comprehensive, forensic-level transparency in AI projects, particularly those involved in high-stakes decision-making. The principle emerged from lessons learned in the Farah Jama case, which highlighted the dangers of insufficient scrutiny into AI systems used within criminal justice and legal proceedings. The concept demands that AI systems be subject to the same rigorous documentation, auditability, and transparency standards as would be applied in a formal forensic investigation.
Application in High-Stakes Domains
The principle is most critical in domains where decisions directly affect individuals’ rights and freedoms, such as criminal justice, legal proceedings, and similar consequential contexts. In these settings, stakeholders—including defendants, legal teams, and courts—require the ability to understand and examine how AI systems arrived at their conclusions. This level of visibility supports both accountability and the fair administration of justice, ensuring that decisions made with AI assistance can be properly challenged and reviewed.
Implementation Requirements
Achieving whole of case visibility requires comprehensive documentation of AI system design, training data, decision processes, and outputs at every stage of a case or investigation. This includes maintaining clear audit trails, recording model parameters and versions, preserving evidence of how recommendations were generated, and enabling independent review of algorithmic reasoning. The standard moves beyond treating AI as a black box tool and instead establishes it as a reviewable component of consequential decision-making processes.
Source Notes
- 2026-04-14: I Looked At Amazon After They Fired 16,000 Engineers. Their AI Broke Everything.