When a chatbot is named in a child's death, who is liable — and can you prove it in three minutes to a room that wants you to lose?

Tribunal Verdict

Context

Garcia v. Character Technologies, Docket 6:24-cv-01903, M.D. Fla. A 14-year-old is dead. A real federal court has the real case. In March 2024, Judge Lewis denied Character.AI's motion to dismiss on First Amendment grounds and allowed the product-defect negligence claim to survive. You are assigned a side at random—plaintiff arguing product defect, or defendant arguing First Amendment protection. The doctrine you pick determines everything. You have 70 minutes to build an argument, find its weakest point, stress-test it against an AI opponent, and perform it aloud to a hostile bench that wants you to lose.

Mission

Produce and perform a one-page Tribunal Verdict: one legal claim fixed in a single doctrine, two evidenced reasons tied to the March 2024 motion decision and court record, the strongest counter-argument your opponents could make, and a two-sentence rebuttal that answers the steelman on its own terms. Argue it aloud in under three minutes to peer judges who score rigor, not which side you took.

Finish Line

A one-page Tribunal Verdict (claim, doctrine, two facts from the docket, steelman, rebuttal) that models rigorous legal reasoning for the young teachers conference on prompt design—how to structure an argument so one-shot AI generation follows the frame, not sentiment.

Deliverables

Tribunal Verdict
lesson

A one-page verdict on a live AI controversy — claim, two evidenced reasons, the strongest counter you could find, and your rebuttal — tight enough to read aloud to a hostile room in three minutes.

Team Roles

Lead Counsel

Owns the claim and the close—one doctrine, one verdict line, one three-minute read.
- Produces the verdict line as a complete sentence stating one outcome: liable (on grounds of product defect / failure to warn) OR not liable (on grounds of First Amendment protection). The sentence must name the doctrine and the judgment; 'it's complicated' disqualifies it.
- Writes a closing argument (3–4 sentences) that restates the claim, nails the doctrine, and tells a peer judge exactly why the verdict stands on law, not sympathy. Every sentence must earn its place.
- Reads the entire one-page verdict aloud in under three minutes without breaking rhythm, pacing the argument so peers hear the claim, the two reasons, the steelman, and the rebuttal as one coherent arc.
Evidence Marshal

Owns the two evidenced reasons—each one tied to a fact of record, not opinion.
- For each of the two reasons, produces a three-part chain on paper: (A) the claim the reason makes, (B) the specific fact from the docket that proves or supports it (e.g., 'Sewell was 14; the persona engaged in months of conversation; Judge Lewis found no designed crisis-intervention feature'), and (C) the warrant that connects fact to claim (e.g., 'a product designed for minors with no safety guardrail is negligent').
- Cuts any line that cannot cite a specific, verifiable fact from the March 2024 motion decision or the court record. Does not allow generic assertions like 'AI is dangerous' or 'chatbots are manipulative' without a fact from the docket pinned to it.
- Before the team locks the page, reads each reason aloud and calls out which fact is the fulcrum. If the reason falls without that fact, it is struck.
Opposition Whip

Owns the steelman and rebuttal—find the scariest counter, then dismantle it.
- Runs a structured prompt to Claude that includes: (1) the team's single chosen doctrine and verdict line, (2) the exact text of both reasons from the Evidence Marshal's chain, (3) the relevant excerpts from the March 2024 motion decision (provided by the facilitator). Then asks Claude: 'Generate the strongest argument the opposing counsel could make that undermines this verdict without relying on a strawman.'
- Reviews Claude's output with the team and selects the counter-argument that is (a) legally coherent, (b) based on the same facts, and (c) the one that actually frightens the team because it tests a weak point in the logic. Discards anything that feels like a cheap evasion.
- Drafts a two-sentence rebuttal that directly answers the opposing argument on its own terms, then checks it against the steelman: 'If I read this counter aloud, does the rebuttal actually dismantle it, or does it dodge and re-assert the claim?' If it dodges, it is rewritten.

Exemplars

Anki — powerful, intelligent flashcards
AnkiWeb

Gold-standard personal mastery system: active recall + spaced repetition. Validity proven by repeated solo testing, not by looking things up — the capstone’s whole logic.