← Evidence Index · denialbydesign.org Pillar 3 · AI Jury Evidence

AI Jury · Independent Verdicts

An independent AI jury was convened to evaluate the merits of the Denial-by-Design class action thesis. Five frontier AI models — across competing labs (Anthropic, Google, OpenAI, xAI, DeepSeek) — were given the named version of the case file and asked to render an independent assessment. Two passes were run for verification.

Why this matters: these are not Anthropic/Claude internal evaluations. They include direct competitors — Google Gemini, OpenAI GPT, xAI Grok, DeepSeek — who have no commercial reason to validate a case that names Anthropic as defendant. The convergence of independent verdicts across competing labs constitutes evidentiary convergent assessment.

9 AI jury responses · click each for the full text

AI modelResponse sizeOpening verdict (first 300 chars)Read
Claude-Sonnet-4.634,401 bytesI am reviewing this brief as a stress-tester, not an advocate. Where the theory is strong, I will say so. Where it is weak or potentially fatal, I will say so plainly. Several of my answers will be unfavorable to the proposed filing. That is the function of this panel....Read full →
Claude-Sonnet-4.6 PASS214,469 bytes**5. Statute of limitations (continued)**...Read full →
DeepSeek-v4-pro10,965 bytes**Q1 · FRCP 12(b)(6) survivability ranking (Microsoft M‑1 through M‑10)**...Read full →
DeepSeek-v4-pro PASS26,909 bytes**Q8 · Statute-of-limitations tolling (completed)**...Read full →
GPT-5.5-pro84 bytesNone...Read full →
GPT-5.5-pro PASS226,822 bytes**Threshold note:** a pro se plaintiff cannot prosecute a Rule 23 class on behalf of absent class members. My analysis assumes class counsel appears before or at filing. If the complaint is filed pro se with class allegations, that alone is a major dismissal/strike risk....Read full →
Gemini-2.5-pro18,892 bytesAs a neutral documentary-evidence reviewer, I have assessed the proposed class-action litigation theory. My analysis is based strictly on the provided brief and my domain knowledge in American, Canadian, and EU law....Read full →
Gemini-2.5-pro PASS25,589 bytesHere are the answers to the remaining questions, followed by the Bottom-Line Verdict....Read full →
Grok-4.36,341 bytesQ1. M-1 (Wiretap Act) and M-2 (Stored Communications Act) rank highest for 12(b)(6) survival because the control-group canary evidence directly implicates pre-delivery scanning without user consent, fitting the statutes' interception and access provisions. M-3 (RICO) and M-9 (tortious deprivation of...Read full →

Source materials

📊 JURY_SUMMARY.json (Pass 1 aggregated) 📊 JURY_SUMMARY_PASS2.json (Pass 2 verification) 📁 Anonymized control jury (same questions, identifying details removed) 📄 Named-vs-anonymized verdict comparison
What you should do next: click any of the response links above. Read what an AI from a competing lab said about Francesco Longo's case when it was given the case file unencumbered. The convergent finding across all five is that the Denial-by-Design thesis is supported by the documentary record.