Ep 2: Case Study 01: Human-in-the-Loop AI & The 2026 AFCON Legal Dispute.
Correction: An earlier draft referred to the 1976 Morocco-Guinea AFCON match as a precedent involving a Moroccan walk-off followed by resumption of play. We no longer rely on that characterization. On further review, we have not identified a primary archival source sufficient to support that account as a settled historical fact, and the reference has therefore been removed from our argument. Our position does not depend on the 1976 analogy and rests instead on the text of the applicable regulations, the structure of the Laws of the Game, and the cited CAS authorities.
What happens when a sports dispute leaves the pitch and enters the courtroom—and the legal strategy is drafted by an AI?
In our first official case study on Code by Reason, we examine a mock legal appeal surrounding the highly controversial 2026 Africa Cup of Nations final. Following a 14-minute stoppage-time walk-off, the Confederation of African Football (CAF) stripped Senegal of their title. We tasked our custom AI agentic framework to review the governing rulebooks and precedents, and it returned an astonishingly detailed, hyper-literal textual defense.
But our goal today is not to declare this AI-generated appeal a flawless legal victory. Instead, we are asking a critical question for builders and researchers: Does this output hold up to professional scrutiny?
This episode is an open invitation to you, our listeners. Whether you have formal legal experience or simply a sharp analytical mind, we want you to evaluate the AI's reasoning. Where are the logical gaps? Can you spot any hallucinations, confident misinformation, or misinterpretations of the law?
By sharing your feedback and pointing out these flaws, you directly help us refine our "human-in-the-loop" methodology and improve the agentic framework. We are testing the hypothesis that AI—when guided by clean data, strict rubrics, and proper oversight—can genuinely contribute to high-level research and discovery as a supportive tool, rather than a standalone replacement for human experts.
Listen in, stress-test the AI's logic, and let us know where the system passes the test—and where it falls short.
Code by Reason is a nonprofit podcast by Project Hamburg Research Inc. (projecthamburg.org), a 501(c)(3) based in Buffalo, NY, and Montreal, QC. We focus on bridging AI solutions with education, law, healthcare, and finance. Learn more and support our research at our website.
