The AI verdict is correct because it doesn’t just rely on isolated data points—it uses a framework that aligns with how science builds consensus. When a claim contradicts the overwhelming body of evidence, the burden isn’t on the evidence to prove itself; it’s on the claimant to explain away the data. The AI’s argument doesn’t just say “HIV causes AIDS”—it shows how denying that link leads to real-world harm, like the South African tragedy. That’s the kind of nuance the moderator should care about: not just facts, but consequences.