The thesis
Public trust in official statistics is collapsing at the same moment our societies need it most. Voters question election tallies. Markets question unemployment rates. Academics cannot reproduce social-science studies because the underlying micro-data is sealed by law. The crisis is not that the agencies are dishonest — most are not. It is that the public has no way to verify.
zkDB allows official agencies to publish the math while keeping the micro-data sealed.
Where we engage
Verifiable national statistics
The agency commits to its micro-data once. Every monthly or annual release — unemployment, CPI, GDP, demographic estimates — ships with a proof that the headline number is the exact computation of the committed survey responses under the published methodology. The micro-data stays sealed; the math becomes public.
Relevant frameworks: US Title 13 (Census confidentiality) · EU statistical-secrecy regulations · OECD Principles for Official Statistics · UN Fundamental Principles.
Verifiable election tallies
Ballots commit to a sealed cryptographic envelope. The tally is a verifiable query. The proof confirms that every committed ballot was counted, exactly once, under the certified election function. Auditors, observers, candidates, and the public verify the proof — none of them see individual ballots.
This is a politically explosive use case. Engagements in this area are reserved for serious election authorities at the federal, state, or supranational level, with the appropriate legal and procurement framework. We do not work with private election-tech vendors.
Cross-border statistical aggregation
Eurostat-style aggregation across member-state agencies, without member states pooling raw micro-data. Each agency contributes a proof that its committed micro-data, under the agreed methodology, produces the contributed component. The cross-border statistic verifies as the deterministic combination.
Verifiable transparency portals
Public-procurement disclosure, beneficial-ownership registers, sanctions enforcement statistics. Each public claim ships with a proof that the disclosure is derived from the agency's committed source ledger. Investigative journalists, civil-society watchdogs, and academic researchers verify independently.
Public-sector AI accountability
For ministries deploying AI in benefits adjudication, immigration triage, or risk scoring: prove that the model output was correctly derived from the citizen's committed application record, under the published model and policy. Citizens, ombudsmen, and courts gain a verifiable audit surface — without the ministry disclosing model internals or other citizens' data.
How an engagement is shaped
Public-sector engagements typically run through a formal procurement framework (G-Cloud, GSA, EU equivalents, or direct ministerial commission). Our first conversation is always confidential and exploratory. We then jointly scope a feasibility brief that can be presented to the agency's legal, procurement, and parliamentary oversight bodies as needed.
What we typically deliver
- Statistical methodology audit and circuit-translation document
- Commitment-ceremony protocol fit for archival-grade record-keeping
- Reference prover and verifier — with multi-language verifier ports so independent observers can run their own
- Key-custody design with hardware-backed segregation across the agency, its inspector general, and the oversight body
- Public-facing explanatory material for citizens, journalists, and academics
- Training for analysts, statisticians, and the agency's IT and information-security teams
Request a briefing
Reserved for senior officials, agency principals, and procurement leads at national statistical offices, election authorities, ministries, supranational bodies, and central banks. Request a briefing →
