ML-DSA-65 for compliance records

ML-DSA is the module-lattice digital signature algorithm standardized by NIST in FIPS 204, derived from the CRYSTALS-Dilithium submission. ML-DSA-65 is its middle parameter set, targeting a security category comparable to AES-192.

Its security rests on the hardness of lattice problems rather than on integer factorisation or discrete logarithms, which is the property that matters: the mathematical structures a large quantum computer would attack are not the ones ML-DSA depends on.

How it compares in practice

• Signature size. ML-DSA-65 signatures are substantially larger than ECDSA P-256 signatures, measured in kilobytes rather than tens of bytes.
• Key size. Public keys are similarly larger.
• Speed. Signing and verification are fast, and for record-keeping workloads the difference is not an operational constraint.
• Maturity. It is a published federal standard with multiple independent implementations, not a research proposal.

For high-frequency transport security, the size difference is a real engineering consideration. For compliance records written at human decision rates and retained for years, it is close to irrelevant.

Why evidence is the natural first application

Post-quantum migration is usually framed around confidentiality: data captured today, decrypted later. Signatures invert the exposure. A record signed today is intended to be believed for a decade, and the adversary chooses when to attack. If the signature algorithm becomes forgeable during the retention period, the record's guarantee evaporates quietly, with nobody watching for the moment it stops being true.

That makes long-retention evidence the place where post-quantum signing has the clearest justification today, well ahead of the general migration timeline most organisations are on.

Signing is not the whole answer

A post-quantum signature proves integrity and origin. It does not establish time. A signature can be produced whenever the key holder chooses, so timestamping still has to come from somewhere the record's owner does not control. Rubric pairs ML-DSA-65 signing with anchoring to a public ledger, so the record carries both properties.

Coexistence with existing standards

Adopting post-quantum signing internally does not require abandoning ecosystems built on classical cryptography. Rubric's documents carry C2PA Content Credentials using the conventional algorithms that specification requires, wrapped around the ML-DSA-65 attestation layer underneath.