# Deterministic and Generative System Boundaries

## Status

Current repository artifact for IARPG-OPS-2 2.0.18-wip. Review status: **reviewed with limitations for scenario literacy and public explanation**. The preserved source file remains individually addressable as provenance. This report is not an official IARPA publication.

## Purpose

Use the supplied architecture report to strengthen IARPG content about deterministic state, generative explanation, API abstraction, validation, and failure handling.

## Scope

This canonical report covers the user-supplied file `IARPA Simulation Game Architecture.md`. It is authoritative for repository provenance, the bounded synthesis below, and IARPG content-routing decisions. It does not establish official agency affiliation, scientific model validity, public-health guidance, weapons effects, real-world geopolitical prediction, or a current software implementation.

## Executive Summary

The source proposes a 3D planetary interface, a PHP backend, domain-driven design, and AI adapters for ecological and geopolitical events. IARPG retains the separation between authoritative rules and generative proposals, the value of adapters and structured outputs, and the need for graceful failure. It does not adopt the proposed platform, external AI dependency, or fictional agency affiliation.

## Evidence Reviewed

- [Preserved source file](../../source-files/planetary-systems-and-scenario-modeling/IARPA%20Simulation%20Game%20Architecture.md)
- Source collection: `planetary-systems-and-scenario-modeling`
- Source SHA-256: `33bb34848eed76943e1fc834ba937d2be5f6639b970de0e6d3643b6d29003fa9`
- Source identity: user-supplied Markdown described as IARPA-related material; not authenticated as an official IARPA publication.
- [IARPA mission](https://www.iarpa.gov/who-we-are/history/our-mission) — official agency identity, research mission, and non-operational boundary (accessed 2026-07-22).
- [IARPA research programs](https://www.iarpa.gov/research-programs) — official program directory (accessed 2026-07-22).
- [GAO infectious disease modeling practices](https://www.gao.gov/assets/gao-20-372.pdf) — communication, model description, verification, and validation (accessed 2026-07-22).
- [NIST digital-twin credibility work](https://www.nist.gov/programs-projects/digital-twins-advanced-manufacturing) — verification, validation, and uncertainty considerations (accessed 2026-07-22).
- [IARPA COSMIC](https://www.iarpa.gov/research-programs/cosmic) — official geospatial program description (accessed 2026-07-22).
- [IARPA HAYSTAC](https://www.iarpa.gov/research-programs/haystac) — official movement-modeling program description and privacy framing (accessed 2026-07-22).
- [IARPA BENGAL](https://www.iarpa.gov/research-programs/bengal) — official LLM threat and resilience program description (accessed 2026-07-22).
- [IARPA SINTRA](https://www.iarpa.gov/research-programs/sintra) — official space-debris research description (accessed 2026-07-22).

## Reviewed Synthesis

### Publication Decision

Retain the source as preserved research evidence. Reuse only the bounded statements in this section and the repository-authored findings below. Do not treat the original title, acronym expansion, code, quantitative claims, named technology stack, or scenario outputs as current fact without a new review.

### Claim Dispositions

| Claim ID | Topic | Disposition | Current bounded statement |
|---|---|---|---|
| `PSR-OPS2-218-05-01` | authority boundary | **retained-method-guidance** | Keep deterministic records authoritative and treat generated text or events as proposals. |
| `PSR-OPS2-218-05-02` | API abstraction | **retained-method-guidance** | Use provider-neutral interfaces, validation, provenance, and explicit fallback behavior when discussing AI architecture. |
| `PSR-OPS2-218-05-03` | external AI deployment | **rejected-for-public-reuse** | Do not imply that IARPG currently calls OpenAI or any other model provider from the public site. |
| `PSR-OPS2-218-05-04` | IARPA affiliation | **attribution-corrected** | Do not present the fictional atlas concept as an IARPA product, endorsement, or official program. |
| `PSR-OPS2-218-05-05` | program descriptions | **requires-current-verification** | Verify current IARPA program names and purposes only against official agency pages. |
| `PSR-OPS2-218-05-06` | geopolitical resolution | **safety-restricted** | Do not use generative systems to make real-world geopolitical judgments, targeting decisions, or automated sanctions. |

### Attribution, Fairness, and Safety Boundary

IARPA is the **Intelligence Advanced Research Projects Activity**, a U.S. research organization that invests in high-risk/high-payoff research for Intelligence Community challenges and is not itself operationally focused. The phrase “Integrated Artificial Reality Planetary Atlas” is a fictional construction in the supplied material and must never be presented as the agency’s name, product, endorsement, or program. Apply equal evidence standards, disclose uncertainty, include rights and affected-community consequences, and keep all reuse non-actionable.

### Reuse Rule

Use the smallest applicable bounded statement above. Re-check current agency programs, laws, software, vendors, performance, market data, and scientific claims against current primary or authoritative sources before public factual reuse.

## Findings

- Deterministic state should remain the sole authority for permissions, transitions, evidence, completion, and version history.
- Generative systems should receive bounded context and return schema-constrained proposals that can be accepted, revised, or rejected.
- Adapters reduce coupling to one provider and make failure, provenance, cost, and privacy policies inspectable.
- Fallback behavior should preserve the publication or scenario experience when an external model is unavailable.
- For current IARPG content, these are explanatory architecture principles—not a claim that a production AI simulation service exists.

## Decisions or Recommendations

- Route public readers to the [Scenario Modeling and Systems Thinking](/scenario-modeling) page rather than to implementation snippets in the preserved source.
- Label every model family with purpose, inputs, assumptions, state variables, outputs, uncertainty, verification, validation, affected communities, and prohibited interpretations.
- Prefer side-by-side scenario comparison, reversible state history, and accessible tables over spectacular but unexplained visualization.
- Keep real IARPA references confined to accurate, sourced descriptions of the agency and its public research programs.
- Preserve the original source unchanged in the source archive, but do not duplicate unsafe or misleading implementation detail into current public guidance.

## Risks and Limitations

- The preserved source contains unsupported, time-sensitive, implementation-specific, and sometimes hazardous detail.
- This review is editorial and architectural; it is not scientific peer review, model accreditation, public-health review, legal review, weapons-effects validation, or security certification.
- Official program pages can change after the research cutoff; future releases must re-check time-sensitive references.
- A scenario can improve reasoning while still embedding bad assumptions. Visible structure does not prove real-world validity.

## Validation Performed

- Verified the preserved source SHA-256 and one-to-one canonical mapping.
- Compared agency identity and selected program descriptions against official IARPA pages on 2026-07-22.
- Compared model-credibility language with GAO and NIST guidance.
- Reviewed the synthesis for attribution, safety, accessibility, international fairness, uncertainty, and non-prediction boundaries.
- Verified repository-relative links and required report sections through the local memory and source-integrity validators.

## Memory References

- [`.uai/planetary-systems.uai`](../../../.uai/planetary-systems.uai)
- [`.uai/architecture.uai`](../../../.uai/architecture.uai)
- [`.uai/ai-source-validation.uai`](../../../.uai/ai-source-validation.uai)
- [`.uai/simulation.uai`](../../../.uai/simulation.uai)

## Related Durable Documents

- [Planetary Systems Research Integration and Scenario Modeling Report](planetary-systems-research-integration-and-scenario-modeling-report.md#findings)
- [Rapid Deep Research Methodology and Guidance](rapid-deep-research-methodology-and-guidance.md#findings)
- [Equal Standards Across Unequal Records](equal-standards-across-unequal-records-international-institutional-fairness-rights-and-evidence-audit.md#mandatory-analytical-distinctions-and-rights-safeguards)

## Supersession Status

Current for IARPG-OPS-2 2.0.18-wip until superseded by a later reviewed synthesis. The preserved source remains archival provenance and does not override active `.uai` memory or verified implementation.