⚠️ Draft placeholder. This document is pending review by legal counsel and is not yet binding.

Engineering and Safety-Critical Use Disclaimer

Version: Draft (pending counsel review) Last updated: 2026-06-14 Effective date: [To be set on publication]


This Engineering and Safety-Critical Use Disclaimer is incorporated into the Terms of Service. It clarifies the limitations and intended use of Artifacts generated by Blueprint.

In plain language: Blueprint is a powerful drafting assistant. It is not a licensed Professional Engineer. It is not certified for safety-critical work. Every Artifact must be independently reviewed by a qualified human before it is relied on for any real-world decision. Failure to do so is at your own risk.


1. What Blueprint is

Blueprint is a software tool that uses large language models to draft systems-engineering artifacts (requirements, decomposition, design, risk register, gate-readiness packages) from a Brief you supply. The drafts are aligned to the NASA Systems Engineering Handbook (NASA SP-2016-6105 Rev 2) and related public standards.

Blueprint is not:

NASA SP-2016-6105 and related standards are referenced because they describe the artifact structures Blueprint targets. NASA does not endorse, sponsor, or certify Blueprint, and Blueprint is not affiliated with NASA.


2. Known AI limitations

Artifacts are produced by large language models. LLMs are capable of producing high-quality drafts but are subject to known failure modes that you must account for:

Failure modeWhat it looks likeMitigation
HallucinationPlausible-sounding but factually wrong contentIndependent verification of every factual claim
Fabrication of citationsReference to standards, papers, or data that don't exist or don't say what's quotedVerify every citation in primary source
Numerical errorsIncorrect tolerances, wrong units, miscalculated valuesIndependent calculation by qualified engineer
Loss of context across stagesEarlier sections inconsistent with later sections after multi-stage pipelineEnd-to-end consistency review
Outdated knowledgeStandards updated since model training cutoff not reflectedConfirm currency of standards referenced
Reasoning gapsMissing edge cases, failure modes, or boundary conditions a senior engineer would considerSenior engineer review

We provide in-product features (citation verification, multi-judge review, sensitivity tier selection) to mitigate these failure modes, but mitigation is not elimination. The product is intended to accelerate qualified human engineering, not to replace it.


3. Safety-critical and certification-bearing use

You agree not to rely on any Artifact, without independent qualified human review and verification, for any of the following purposes:

3.1 Safety-of-life or safety-of-flight

Decisions where failure of the design could result in death, injury, or significant physical harm. Examples (non-exhaustive):

3.2 Certification or regulatory compliance evidence

Documentation submitted to a regulatory body as part of a certification or approval package. Examples:

3.3 Final design decisions for high-consequence systems

Designs where failure would result in significant property loss, environmental damage, or critical infrastructure impact.

3.4 Legal, financial, or compliance opinions

Any Artifact relied upon as a legal opinion, financial analysis, or compliance attestation. Blueprint is not a law firm, accounting firm, or compliance auditor.


4. What "independent qualified human review" means

For purposes of §3, "independent qualified human review and verification" means review by a human professional with appropriate licensure, experience, or certification for the specific decision being made, who:

A non-licensed engineer can use Blueprint Artifacts as a drafting aid; relying on them for certification-bearing purposes is a distinct decision requiring distinct accountability.


5. Sensitivity tiers and applicability

This disclaimer applies at every sensitivity tier, regardless of whether the underlying LLM provider is Standard, Confidential, or Controlled (when available). The sensitivity tier addresses data-handling, not output-correctness. AI Output limitations apply uniformly.


6. Acknowledgment

By using the Service, you acknowledge that you have read this Disclaimer, understand the limitations described, and accept full responsibility for any reliance on AI Output beyond the limitations described.


7. Contact

Questions about this Disclaimer:

NeuronKite LLC Attn: Legal 4539 N 22nd St #6544, Phoenix, AZ 85016 [email protected]