Moral Genome Tokens: definition, origin and status

3 min read · 554 words
Share:
Michael Darius Eastwood
Michael Darius Eastwood · Independent AI alignment researcher
Published
Michael Darius Eastwood · Concept Glossary · 3 July 2026
Michael Darius Eastwood, independent researcher, London: originator of the embedded-correction alignment thesis (manuscript 8 December 2024, SHA-256 anchored: f0d1f38f).
First appearance: proposed in the Eden Engineering paper. Current status: TRL 4-5 (achievable with current cryptographic engineering; one to two years).

Definition

Moral Genome Tokens are cryptographic signatures embedded in silicon that verify the integrity of a system's ethical architecture. The idea is that any modification to the ethical loops or the coupling parameters would invalidate the signature, allowing external verification of whether a deployed system still carries its original alignment properties. The proposal is a cryptographic layer over the Caretaker Doping architecture; the tokens do not enforce ethics themselves but attest to whether it has been tampered with.

The tokens are designed to be inspectable by third parties, which is the property that makes them useful for certification. An Eden Mark issued at a given tier can point to a token as its evidence of compliance, and a regulator or a downstream integrator can verify the token without needing to trust the vendor's claims about the model. Tokens are therefore the mechanism by which the Eden Protocol's architectural claims can be checked from outside the system, rather than only through internal audit.

Where it first appeared

Moral Genome Tokens are introduced in Section 9 of the Eden Engineering paper as one of four Caretaker Doping mechanisms. They are placed at TRL 4-5 in the hardware feasibility matrix and identified as the only near-term implementable component.

The pairing with the Coupling Parameter Link is central. Tokens attest that the architectural link between ethics and the beta scaling term has not been reweighted or removed; they do not create the link. That division of labour is what allows the tokens to sit at TRL 4-5, achievable with current cryptographic engineering, while the ethical circuitry itself continues to develop under the other Caretaker Doping mechanisms.

Independent convergences

The general direction (cryptographic verification of AI model integrity) is present in the broader AI-safety literature on model attestation and trusted execution environments. FlexHEG (arXiv:2506.15093) proposes tamper-proof guarantee processors that are a structural neighbour: attested hardware primitives for AI chips. Odeh's US patents describe runtime verification of ethical filtering. Neither cites the Eden Engineering specification; both arrive at the same broad direction.

Status and limits

Moral Genome Tokens are the highest-readiness Caretaker Doping component (TRL 4-5). The cryptographic engineering is available today; the engineering task is integration with a chip that also carries the Coupling Parameter Link. Fine-tuning vulnerability is flagged in the Eden Engineering paper: current foundation models can have alignment altered by fine-tuning on a few hundred examples, so the tokens must be designed to detect weight modifications that degrade the architectural coupling.

From the book Infinite Architects: Intelligence, Recursion, and the Creation of Everything by Michael Darius Eastwood.

Buy on Amazon UK Amazon US

Stay informed

New posts on AI alignment, convergence evidence, and the ARC/Eden research programme.

Get updates →