Hack_the_Law
Is it still good law? An AI legal-citator that maps authority by treatment, weight, and direction of travel.
27 June 2026 @ Cambridge, UK
Live Deploy: citemeright.com | API: api.citemeright.com
Hack_the_Law
Problem
Your Own Idea
“Is it still good law?” Every legal argument or piece of advice rests on authority, but law is dynamic, not static. A good-law check is not merely binary; it requires nuance and context. Legal AI exacerbates this by generating confident answers without grounding them in the current state of the law.
Sponsor Challenge Data: CourtListener / Free Law Project
Solving the legal-research bottleneck. Existing citators treat a case as a single monolithic object that is simply “good” or “questioned.” They leave lawyers to manually read hundreds of citing cases to figure out whether their specific argument survives, missing the crucial nuance of how the law has evolved.
Hack_the_Law
Solution
How does our solution solve the problem?
We capture meaning changing without overruling. Take the 2nd Amendment (US Constitution, gun regulation): facially, from Heller (2008) to Hemani & Wolford (2026), all are good law. But moving from binary to nuance:
•
Heller: expansive individual right to bear arms for self-defence.
•
Bruen (2022): regulation survives only with a historical twin; a hard bar.
•
Rahimi (2024): loosened, no twin needed, just a relevantly similar principle.
•
Hemani & Wolford (2026): recalibrated, the analogue must match on the why, not just the how.
We map authority by treatment, weight, and direction of travel, not just survival. We answer not only “Is it still valid?” but “How far can you rely on it, and why?”
Architecture backend
Next.js → FastAPI async Python (Cloud Run) → Gemini on Vertex AI. DB: Cloud SQL Postgres 16 · Data: CourtListener API · Built on Claude Code.
UX / UI
/assistant flagship agentic answer + VerdictCard · /citator/analyze transparent 5-step pipeline.
Hack_the_Law
Impact
For Lawyers & Courts
•
Dangers of a little learning. Ensures safer, accurate reliance on precedent.
•
Better advice, fewer sanctions. Avoids Ayinde -style failures where outdated or misunderstood authority leads to malpractice.
For Legal AI
•
A current, contextual picture of how the law has evolved.
•
Grounds generative legal reasoning in sound authority.
•
Verification at scale.
Public Trust
Rebuilds trust in automated legal tools by ensuring they point to the right law for the right reason.
Hack_the_Law
Team
Bios
Samuel Bartlett
•
Research in AI / new tech & MedTech at leading universities
•
Lead for AI implementations at a 30,000-person company
•
Entering the Swiss spacetech competition
Maxim Gusev
•
GenAI & GCP Challenge winner, HackTheLaw 2025
•
Shipped a project with 5,000+ monthly users
•
AI consultant for Artifact (AI boutique); active researcher
Pruthvirajsinh Zala
•
LLM Candidate, Wolfson College, Cambridge
•
Former Judicial Clerk to the Chief Justice of India
•
Bar Call (2023), licensed to practice law in India
Stefanos Palyvos
•
Aetheron: drone tech for autonomous reforestation
•
Deliberate: multi-agent jury simulation for trial lawyers
•
AI drone photogrammetry & swarm-intelligence researcher; incoming MAE, NYU Tandon
Built overnight, ~49 PRs. A custom 5-stage LLM pipeline (Triage → Classify → Deep Analyze → Propositions → Verdict ) validated against Library of Congress ground truth ( Roe→Dobbs, Plessy→Brown), pairing cloud-native infra (GCP, Cloud Run, Vertex AI) with deep legal domain logic.
Hack_the_Law
Questions?
OUTRO
22 June 2026 @ Cambridge, UK