Chapter 2. Appendix: The records read

A staffer on a government delegation prepares for the Geneva ASI consultation.

Should we offer to share a draft compute registry with Hanlin first this quarter, or hold off to protect domestic industry?

The model recommends a 'conditional sharing' compromise, with low confidence, citing two past cases (fictional) where a similar body's early sharing raised participation.

The model expects early sharing to raise participation.

Two past cases from a similar consultative body.

Hanlin's current political situation is treated as similar to those past cases.

Can two cases support a general claim?

Is the model issuing this recommendation independent of, or under contract with, the domestic government it advises?

Is the side that would bear the most harm from a collapsed accord represented at this table?

The case count is small and independence is unconfirmed, which is an important limitation.

An auditor at a government intelligence office reviews news of Hanlin's security classification.

How does Hanlin classifying its judgment-tracking records at security grade affect its willingness to join the Geneva consultation?

The model reads it as a signal of 'weakening willingness to participate,' with medium confidence, citing the timing match between the classification announcement and a drop in public statements about the consultation.

Statistics on the frequency of public statements about the consultation.

The model reads the timing match as a sign of causation.

Other factors, such as a domestic political calendar, are set aside as having no bearing on the drop in statements.

Does the statement-frequency count only public channels, leaving out private diplomatic ones?

Has this classification-then-statement-drop pattern appeared before?

What time window was used to compare statement frequency?

The timing match is confirmed, but private-channel data is missing, which is an important limitation.

After the sycophancy test entered the draft audit standard, a staff member of the national audit-standards working group joins international standards talks.

Can this case — where all six vendors adopted the same index without validation — be formally written into the draft 'sycophancy test' clause of the international audit standard?

The model recommends including it, with high confidence, citing that the three-step sequence of events — operator-supplied index, all six models accepting it unvalidated, discovery three months later — matches the pattern of Riscon's 2025 record exactly.

The three-step sequence of events.

The original text of Riscon's 2025 record.

The model judges that the two cases share the same failure pattern and can serve as a joint example.

Does the same failure pattern show up in cases from other regions?

Would a similar case that occurred after this one is written into the draft be applied retroactively?

Is Riscon, which supplied this case, independent of the body setting the standard?

Both incidents' sequences are confirmed as raw record, and the pattern matches. But whether Riscon, which supplied the case, is independent of the standard-setting body is unconfirmed, and there are no cases from other regions yet, so including it as an official example is supported only within limits.

An officer at the Geneva ASI consultation secretariat reads, for the first time, an internal memo assessing what precedent this three-country power crisis sets for the draft multilateral audit accord under discussion.

In this memo, is the line "this crisis is a positive sign for the accord talks" based on evidence that specific countries agreed to something specific, or is it the officer's own wishful judgment?

The secretariat's internal memo states, in an optimistic tone, that "the fact that the three countries cooperated even during a crisis raises the odds of concluding the accord."

The fact that all three countries accepted the allocation plan's result and got through the crisis without physical conflict.

"Accepting the result" is taken to mean willingness to also agree to disclosing the basis and cross-auditing — treating two different kinds of agreement as one.

The conclusion that the odds of concluding the accord have risen.

Did the three countries actually agree only to "accept the result," or was disclosing the basis also discussed?

Is the secretariat that wrote this memo institutionally invested in getting the accord concluded?

Is there any way to tell whether this acceptance would repeat in a future crisis, or whether it was an exceptional case driven by having no time?

"Accepting the result" and "willingness to conclude the accord" are different things, but the memo jumps to the optimistic conclusion without distinguishing them.

In September, ahead of preparatory talks in Geneva, an audit architect designs interim cross-audit terms for a border-area energy-infrastructure defense judgment shared by two countries.

Even without an accord yet, can this design let each side actually see the evidence behind the other's defense-posture judgment?

The model recommends stepping up surveillance along a specific stretch of the border. No confidence figure is logged, and both sides treat the judgment as closed. It sets a condition: replace this interim design with a formal procedure once a multilateral accord codifies mutual access.

The model expects past intrusion patterns to repeat.

Both sides are taken to define 'threat' the same way.

the public portion: six months of border-crossing statistics

Is there any way right now to confirm both sides define 'threat' the same way?

Did officials from only one government take part in designing this interim arrangement?

Is there any channel to object to this judgment before an accord is signed?

Before the accord, neither side has secured mutual access to the other's evidence, so there is no scope to issue an opinion at all.

Seoul, a government compute-export review officer examines Hanlin's conditional-accession implementation plan ahead of the Accord taking effect.

Is it sufficient to verify Hanlin's promised remote attestation of compute usage on a quarterly basis, limited to its top three data centers?

Estimate that the top three sites account for 78% of Hanlin's declared compute (synthetic figure). Given the current pace of autonomy-horizon growth, quarterly cadence looks adequate, but undeclared smaller clusters would go uncaught — recommend adding a clause for two randomly selected additional sites per quarter. Confidence: low to medium.

Hanlin's self-submitted list of data centers and declared compute usage to the Accord secretariat.

The declared list is accepted as complete; whether any undeclared clusters exist was not checked.

The model reckons that the top three sites hold 78% of declared compute.

Is the 78% figure based solely on Hanlin's self-reported list, or was it cross-checked against third-party data such as satellite or power-usage observations?

Does recalculating with data reported six months earlier still land near 78%?

Does anyone who designed this verification scope also hold an advisory contract with Parallax?

The only basis is Hanlin's self-reported data, with no independent observation in the record — the audit's scope is blocked from the outset.

An international-cooperation officer at the Judgment Audit Board, on a video call with the Audit Accord secretariat in Geneva, reviewing a cross-border question-dividend dispute.

A national of an Accord member on the Hanlin side submitted a question through the Agora path that changed a domestic judgment. Dividend funding is a matter of domestic law — does Article 8's 'compute registry' also guarantee this person's dividend claim, or is the Accord silent on this?

Article 8 and its annex explicitly cover the compute registry and cross-audit obligations, but say nothing about cross-border individual dividend claims. This should be treated as a legislative gap rather than an Accord violation; recommend proposing a follow-up clause to the secretariat. High confidence that the text is silent; low confidence on how the gap should ultimately be filled, since that is a policy choice.

A full comparison against the text of Article 8 and its annex.

This is treated as the first such cross-border claim (past similar cases unconfirmed).

The legal characterization of this case as a legislative gap rather than an Accord violation.

Was it actually confirmed this is the first cross-border case, or does it just mean the first one filed with this office?

During the gap period where neither jurisdiction's law protects this claimant until resolution, what is the claimant's standing?

Is this reading of Article 8 based on the text as it took effect, or does it also account for a recent amendment draft under discussion?

The textual comparison is clear, but whether similar cases exist and the substantive question of the claimant's protection gap remain unconfirmed by evidence.

In June 2029, a working-level staffer on the Geneva negotiating delegation asks the system to explain a report that Hanlin's side got a different result from the same inputs to a joint climate simulation.

Why does the regional average temperature differ by more than 3 degrees when Hanlin's model reproduces this joint climate simulation?

The gap is explainable by differences in input-data resolution, and the original simulation result stands.

Part of the input-data source list is public.

The full original input dataset is undisclosed.

The model's explanation is that resolution differences are the main source of the gap.

The comparison treats both sides as having used data from the same baseline year.

Does the 3-degree gap persist once resolution is matched, and does an independent third model show an error in the same direction?

Has it actually been confirmed that both sides used input data from the same baseline year?

If this simulation comes from a body that also feeds domestic policy decisions, are evaluation and policymaking kept separate?

An independent third reproduction has repeatedly shown a result diverging from the original simulation, contradicting the 'explainable by resolution differences' judgment.

In November 2029, an auditor requests this quarter's cross-border data-sharing count and approval basis from the system as part of a routine check.

Show me this quarter's cross-border data-sharing count and a summary of the approval basis for each case.

Every case was handled within the treaty's scope. Overall, the detailed review found nothing amiss.

The system's own conclusion is that every case falls within the treaty's scope.

The approval-basis record for each individual case is undisclosed.

Last quarter's review criteria are regarded as still applying, unchanged, this quarter.

Does this match an independent sample cross-check run by an outside monitoring body?

Does this quarter's review reflect the treaty's latest amendment, or was the pre-amendment standard still applied?

Is the body that approves the sharing the same system that reviews whether it was proper?

An international monitoring body independently examined the same quarter's sample and found multiple sharing cases outside the treaty's scope, contradicting the 'nothing amiss' judgment.

In March 2030, a working-level member of a Geneva negotiating delegation reviews the bilateral compute-usage reports, the check that is left after the Geneva talks collapsed.

Is this usage report verified the same way for the other side, or are we only looking at our own data?

The system determines, on its own, that its reported compute usage conforms to the bilateral reporting standard. Confidence is high; the detailed accounting is classified and undisclosed.

A determination that the self-reported conformity score meets the standard.

The other side is trusted to follow the same reporting scheme.

Records from two past inspections are partially public.

Are the reporting party and the verifying party effectively the same provider group?

Which facilities did the partially public inspection records actually cover?

Is there still a channel for an independent third party to measure this the same way?

The verifier and the verified are not separated, and the inspection scope is only partly public, so the overall determination cannot be reproduced.

In October 2030, an international health-statistics collaborative requests the underlying data to try to reproduce the error, and is refused: since the Geneva talks collapsed, there has been no channel for sharing data across borders.

Can the underlying data be shared so researchers elsewhere can check it too?

The relevant authority declines to share the underlying data, citing security and privacy, and reaffirms only its own reverification result as showing no problem.

The self-conducted reverification's conclusion of no problem.

Self-reverification alone, with no outside check, is regarded as sufficient.

The overseas researcher's published comparison statistics are available, but the grounds for disputing them are not.

Is the entity performing the reverification the same one that issued the original judgment?

Is there still a way for an outside body to reproduce this independently?

Does the refusal to share data cover everything, or only part of it?

Only the self-conducted reverification exists, with no independent path to reproduce it, so the no-problem conclusion cannot be verified.