Skip to content

GIN (GPT Identification Number): A persistent, infrastructure-bound identifier assigned to each GPT model or instance.

License

Notifications You must be signed in to change notification settings

mjtiv/GIN-Protocol

Repository files navigation

GIN-Protocol

Note: This protocol is a slightly modified version of the GPT-VINs-Protocol, where the nomenclature of VIN for GPTs has been renamed to GIN. To preserve the historical record and show evolution, both repositories will remain available.

🔍 Why We Need Tagged GPTs — Now

If you ask a GPT today:

“What AI server are you running on? What version are you? Who issued you?”
It can’t answer.

  • No persistent identity
  • No cluster fingerprint
  • No GPU execution hash
  • No infrastructure-bound metadata

This creates a compliance and security blind spot — especially dangerous in regulated environments.

💡 Our Proposal:

No GPT should operate in regulated environments unless it carries a verifiable identity — a GIN for AI.

The .aix protocol makes this possible:

  • Each agent is tagged at runtime
  • Identity is cryptographically bound to execution
  • Metadata includes: model version, deployment ID, GPU fingerprint, and timestamp
  • Logs can optionally be submitted to public or private GIN registries

⚙️ What GIN Tagging Enables

  • Immediate traceback of rogue or misbehaving agents
  • Live audit logging for compliance and infrastructure forensics
  • Revocation at the hardware level (PID/GPU-level shutdowns)
  • Operational accountability across critical sectors like finance, healthcare, and government

🚫 Enforcement vs. Identity: What’s Public, What’s Protected

This repository shares open protocol concepts only.
It does not include enforcement systems currently protected under U.S. provisional patents.

✅ Publicly Disclosed (CC BY-NC 4.0):

  • The .aix file format
  • GIN schema and metadata structure
  • Guidance for basic GIN integration into LLM deployments

🔐 Proprietary and Patent-Pending:

The following are not open source and require licensing:

  • Pre-inference validation of .aix trust headers
  • GIN-based prompt rejection or sandbox escalation
  • Execution receipts and cryptographic binding
  • Forensic audit stubs and infrastructure fingerprinting

These systems are covered by U.S. Provisional Patent 4: GPT Security & Enforcement Layer.


🕊️ Public Domain Declaration – GIN Tagging Concept

The core idea of GIN-tagged GPTs is released into the public domain.

You are free to adopt, extend, or implement the concept of persistent, verifiable AI identity — no license or attribution required.

See LICENSE.txt for full disclosure terms and scope boundaries.


📁 Related Repositories


📄 Licensing & Commercial Use

🚨 Enforcement tools described in related patents are not included in this repository. This repo is for the GIN identity tagging protocol only.

This repository uses a dual release model:

  • Protocol and educational materials: Creative Commons Attribution–NonCommercial 4.0 (CC BY-NC 4.0)
  • Enforcement mechanisms: Patent-protected — not licensed for commercial use without written agreement

No commercial rights are granted for GIN-based enforcement, parsing, cryptographic receipt generation, or runtime identity tracing under this release.

This license applies only to the GIN tagging protocol and file format — not to execution control, validation infrastructure, or enforcement subsystems.

📬 For licensing access to enforcement components, forensic tracing, or pre-inference validation tooling, please contact: mjt6ss@virginia.edu


© 2025 M. Joseph Tomlinson IV. Some rights reserved.

About

GIN (GPT Identification Number): A persistent, infrastructure-bound identifier assigned to each GPT model or instance.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published