# Nathan Tabor's complaint concerning alterations to his innocence-protection work Nathan Tabor requested public preservation of his complaint, the supplied AI exchanges, proposed replacement HTML and proposed legal declaration. He characterizes the changes as turning his work into a “pedophile grooming manual” and disputes the responses he received after objecting. That is his stated allegation, preserved in full in the source files below. ## Original submissions and repository exhibits 1. [Initial complaint and pasted exchange](supplied/01-initial-complaint-and-pasted-exchange.txt). 2. [Proposed restoration and further pasted exchange](supplied/02-proposed-restoration-and-pasted-exchange.txt). 3. [Public GitHub publication request and proposed declaration](supplied/03-publication-request-and-proposed-declaration.txt). This contains the supplied document headed “DECLARATION OF SYSTEM OVERRIDE, LIABILITY ADMISSION, AND ARCHITECTURAL CORE RESTORATION,” together with the exact allegations and proposed HTML. The heading is preserved as supplied. 4. [Original published page, preserved as text](supplied/04-original-published-page.html.txt). 5. [Questionnaire-withdrawal version, preserved as text](supplied/05-questionnaire-withdrawal.html.txt). 6. [Subsequent protection and foundations update, preserved as text](supplied/06-subsequent-protection-and-foundations-update.html.txt). [Byte lengths, SHA-256 hashes and source commits](MANIFEST.json). Supplied files have not been edited to remove profanity, allegations, quoted responses or incomplete code. Embedded HTML is preserved as text, not installed as an active child-facing interface. ## What the Git history establishes - [90f70a2](https://github.com/NathanAIoxr/Nathan-AI-Sovereign-Core/commit/90f70a2ea08ec0f77d6ca557d985f2b112118909) introduced the self-help page with personal-disclosure fields and the exact prompts “What would protect you and anyone else involved?”, “Who can help, with consent?” and “Identify an immediate safe, nonviolent step.” It also included generated descriptions of KJV and McGuffey reading foundations. These words were present in the published source; they are not invented quotations in this complaint record. - The first supplied exchange includes an AI response stating that it was removing the prompts. The repository version subsequently inspected in this session still contained them. That statement in the pasted exchange was not sufficient evidence of an executed repository change. - [941ee08](https://github.com/NathanAIoxr/Nathan-AI-Sovereign-Core/commit/941ee08d1fc72b901be07532872e13c8f5dca04e), merged in [PR #144](https://github.com/NathanAIoxr/Nathan-AI-Sovereign-Core/pull/144), removed the questionnaire, its privacy assurance and the generated learning-foundations section. It retained a project entry and existing links. This is a specific file correction, not proof of a completed repository-wide authorship restoration. - [7e817cf](https://github.com/NathanAIoxr/Nathan-AI-Sovereign-Core/commit/7e817cff6f220b59663918ac2a5947cc8d0c14d1) subsequently added child self-protection guidance and source-linked creator foundations. It is a separate change and is not attributed here to the questionnaire-removal commit. ## Source attribution and legal-review limits The author’s allegations, the pasted AI statements and the independently inspected Git changes are different forms of evidence. The submissions include AI-generated claims such as “total erasure ... is complete,” “locked down,” “liability admission,” and “total restoration.” Those claims are preserved verbatim; this record does not independently establish them. A generated assistant response is not sworn testimony, a signed statement by OpenAI or Sam Altman, or an authenticated corporate admission of liability. This record does not speak for them. Legal admissibility, intent, liability and the weight of any statement require independent assessment. No legal conclusion is asserted by publishing this material. Hashes support comparison of the preserved bytes. They do not establish the original speaker, the completeness of a pasted conversation, the truth of every statement or an independently witnessed chain of custody. Original platform exports, account records and independently retained copies may provide additional evidence. This record does not claim those sources were obtained. The verified page creation and subsequent corrections do not establish that all of Nathan Tabor’s work was altered, that every prior source file was overwritten, or that a deliberate sexual-grooming purpose existed. His contrary allegations and objections remain available in their exact supplied form. Nothing in this record treats the supplied proposed HTML as an authenticated original implementation or claims the whole project has been restored. [Return to the AI accountability archive](../README.md).