Provenance

The Iron Gate Doctrine

AI tools are used throughout this research programme as computational instruments. They are not co-authors. They are not replacements. They enable the production of the researcher’s directing – executing tasks at the researcher’s instruction and under the researcher’s control.

The Iron Gate Doctrine establishes that if every human input, every machine processing step, and every output is forensically documented with cryptographic verification, the question of authorship resolves itself. The provenance chain is the proof.

“The saws did the cutting. But the carpenter built the house.”

Forensic Correlation Methodology

Every artefact produced in this research programme is documented through a forensic correlation chain:

1. Human input – timestamped direction, decision, or creative instruction from the researcher
2. Machine processing – the AI tool’s response, transformation, or output
3. Output – the resulting artefact (text, analysis, code, document)
4. Cryptographic seal – SHA-256 hash of the output, timestamped and logged

This chain is machine-readable, independently verifiable, and immutable once sealed.

Open-Source Tool

forensic_correlation_log.py – A computational provenance module that documents human-AI collaboration sessions with timestamped, hashed, machine-readable entries.

The tool is open-source and available for independent verification.

Verification

To verify any published artefact against its provenance hash:

shasum -a 256 [filename]
# Compare output to the published hash

If the hashes match, the artefact is authentic and unmodified since the seal date.

Verified Artefacts

This table will be updated as artefacts are published and sealed.

Standards

• Hashing: SHA-256
• Transport encryption: TLS 1.3
• Storage encryption: AES-256 at rest
• Logging: Immutable JSONL with monthly hash sealing
• Audit: Annual review per UK GDPR compliance framework

MachenTagar | uk@machentagar.ca