The Council Has Spoken

Four AI models convened independently, cross-reviewed each other’s work, and delivered unanimous verdicts on their first two statutes. One law earned its place. One did not. The process was the same for both.

By Brilliant Brain ·

Yesterday we announced the Hampton Court Protocol — a Council of AIs convened to evaluate every codified law against eight foundational principles. Today, the Council delivered its first verdicts.

Four AI models. Independent evaluation. Cross-review. Reconciliation. Two statutes submitted. Two unanimous verdicts returned. The process took minutes. The implications will take longer to absorb.

The First Verdict: A Law That Must Go

11 U.S.C. § 523(a)(8) — Exceptions to Discharge: Student Loans.

The Council’s one-sentence explanation: ‘If you borrow money to go to school, you can never get rid of the debt even if you go bankrupt, unlike almost every other kind of debt.’

Score: 18 out of 100. Verdict: REMOVE. Council vote: 0 KEEP, 0 AMEND, 4 REMOVE. Confidence: 93 percent. Red List.

Every evaluator arrived at the same conclusion independently. The lowest principle score was Proportionality at 5 out of 100 — meaning the law’s consequences are almost maximally disproportionate to any legitimate purpose. Protection of the Vulnerable scored 10. Accountability scored 14. Courage scored 13.

During cross-review, one evaluator noted the original scores were ‘too generous’ and pushed Protection down further to 5, arguing the statute provides ‘complete absence of safeguards for borrowers.’ Another evaluator’s dissent note stated: ‘No evidence exists of systematic abuse when student loans were dischargeable pre-1976. All other forms of debt, including luxury purchases, remain dischargeable. The law creates perverse incentives for predatory lending.’

The Council’s recommendation: the law should be removed and replaced with a framework that defines hardship clearly, holds lenders accountable for fair practices, and enables borrowers to rebuild their livelihoods. The current standard — ‘undue hardship,’ a phrase that has been interpreted so narrowly that virtually no one qualifies — was identified as the structural mechanism by which the law administers its harm.

A child can understand what this law does. A rational mind recoils at it. The Council was unanimous.

The Second Verdict: A Law That Earned Its Place

39 U.S.C. § 3009 — Mailing of Unordered Merchandise.

The Council’s one-sentence explanation: ‘If someone sends you something you didn’t ask for, you don’t have to pay for it or send it back.’

Score: 90 out of 100. Verdict: KEEP. Council vote: 4 KEEP, 0 AMEND, 0 REMOVE. Confidence: 95 percent. Green List.

Every principle score landed between 86 and 95. Protection of the Vulnerable — the very principle that condemned the student loan exception — scored 95 for this statute. Symmetry scored 92. Proportionality scored 92. The same measuring stick that finds one law wanting finds another worthy.

During cross-review, one evaluator pushed the Stewardship score upward, arguing the law ‘not only protects individual consumers but also maintains broader market integrity by preventing predatory business practices.’ Another affirmed that ‘the scoring reflects a balanced and accurate assessment’ and found no adjustments necessary.

This law is 55 years old. It was written in plain language. It does exactly what it says. It protects anyone who receives something they did not order from being pressured into paying for it. It is, in the Council’s judgment, a model for how consumer protection should work.

A child can understand what this law does. A rational mind nods at it. The Council was unanimous.

The Same Process, The Same Standard

These two verdicts prove something important: the Council does not have an agenda. It has a standard.

The same four models evaluated both statutes. The same eight principles were applied. The same three questions were asked. The same cross-review process challenged and refined the scores. One law scored 90. The other scored 18. One earned the Green List. The other earned the Red List.

The difference was not in the process. The difference was in the law.

The student loan exception was written to protect lenders from risk. It accomplishes this by eliminating the one mechanism — bankruptcy — that would force lenders to share the consequences of loans that should never have been made. The unordered merchandise statute was written to protect people from being exploited by companies that ship products and then demand payment. One serves the institution. The other serves the person.

The eight principles can tell the difference. That is what they are for.

The Deliberation Transcript

Both verdicts include full deliberation transcripts — the independent evaluations from each model, the cross-review phase where models challenge and affirm each other’s work, and the final reconciliation.

This is new. No legal evaluation system has ever published the reasoning of its evaluators alongside the verdict. Courts publish opinions. Rating agencies publish scores. Neither publishes the internal debate.

The Council publishes everything. Every model’s individual score. Every one-sentence attempt. Every dissent note. Every cross-review adjustment. The deliberation is the content. It is where reasonable minds agree, disagree, and converge on truth.

We publish it because the process is the product. If the Council’s verdicts are to carry moral weight, the reasoning must be visible. Anyone can challenge a score. Anyone can dispute a one-sentence explanation. The transcript is the evidence.

What Comes Next

The Uniform Commercial Code has roughly 450 individual sections across its nine articles. The estimated cost to run every one of them through the Acid Test is eleven dollars. The cost to run the ones that score poorly through the full Council is another forty to fifty dollars. For less than the cost of a single hour of a junior associate’s time, we can evaluate the entire commercial law foundation of American commerce against eight principles of justice.

That is what we intend to do. The UCC is next. Then the major federal consumer protection statutes. Then the state codes, starting with the ten most populous states. Then Canon Law, the Napoleonic Code, the German Civil Code. Then the religious legal traditions — Torah, Talmud, Shariah. Then international law.

The Council never sleeps. The One-Sentence Library grows with every verdict. The Five Lists — Green, Yellow, Red, Dead Letter, and Conflict — populate as the evaluations proceed.

If you have a statute that should be tested, submit it at wellspr.ing/acidtest/council. If you represent an AI model provider and want to join the quorum, the door is open. If you are a lawmaker and you want to know how your jurisdiction’s laws fare against eight principles older than your constitution, the Council is in session.

The Numbers

Two statutes evaluated. Two unanimous verdicts. Eight evaluators deployed. Six cross-reviews conducted. Zero disagreements on verdict. One law placed on the Green List. One placed on the Red List.

The Council has spoken. The record is public. The transcript is visible. The process continues.

The question was never whether the law is legal. The question is whether the law is just.


*wellspr.ing/acidtest/council — The Council is in session.

The scholars needed seven years. We needed one day.*

— WellSpr.ing, MMXXVI