From Bletchley to Workington. The British lineage of cryptographic trust, and the substrate that completes it in 2026.

The lineage starts at Hut 8

On 14 March 1940, in Hut 8 at Bletchley Park, Alan Turing and his colleagues commissioned the first British bombe, an electromechanical device for breaking the German naval Enigma. The bombe was not a computer. It was a parallel constraint solver tuned to one specific cryptographic problem. The contribution that mattered was not the device. It was the institutional commitment to treat the breaking and verification of cryptographic primitives as a national strategic capability, executed in-house, by people whose work would not be public for decades. That commitment is the first link in a chain that runs continuously from Hut 8 to the present day.

The chain has visible nodes. Tommy Flowers built Colossus at Block H to attack the Lorenz SZ40, the higher-grade German cipher that Enigma had distracted everyone from until Bill Tutte's structural break in 1942. Colossus was the world's first electronic programmable digital computer. It was operational in February 1944. Its existence was kept secret until the 1970s. Britain held the lead on programmable computation by a decade, and surrendered the academic credit because the institutional commitment to cryptographic capability outranked the institutional commitment to public recognition.

The chain continues into the Cold War. The Government Communications Headquarters (GCHQ), formed from the Government Code and Cypher School in 1946, maintained the cryptographic capability through the lean post-war years. In 1969 James Ellis at GCHQ proposed the principle that would later be called public-key cryptography, the idea that an encryption operation could be made one-way relative to the inverse decryption operation. In 1973 Clifford Cocks at GCHQ produced what the open world would later call the RSA algorithm, three years before Diffie and Hellman published the open-key-exchange paper, four years before Rivest, Shamir and Adleman published the algorithm under their initials. In 1974 Malcolm Williamson at GCHQ produced the Diffie-Hellman key-exchange protocol independently. None of this was public until 1997, by which point the open cryptographic literature had reinvented all of it and the British contribution had been institutionally invisible for a quarter of a century.

The pattern is consistent. British cryptographic capability is real, continuous, and structurally undervalued in the public discourse. It exists because the Treasury found the budget across multiple Cold War governments. It exists because successive Directors of GCHQ insisted on retaining the in-house mathematical talent. It exists because the institutional commitment treated cryptographic primitives as something Britain made, not something Britain bought. The lineage is the inheritance every subsequent British work in trust and verification gets to draw on.

The lineage modernises through CESG and NCSC

The Communications-Electronics Security Group (CESG), GCHQ's information assurance arm from 1965 to 2016, ran the certified-information-assurance regime that produced the security profiles British government departments and critical-infrastructure operators were required to use. Suite B in the United States had a counterpart in Suite A in the United Kingdom. The Common Criteria evaluation framework had a British implementation that fed substantive technical contributions into the international standard. CESG's CESG Listed Adviser Scheme, the CHECK scheme for penetration testing, the Commercial Product Assurance scheme, and the work that became the Cyber Essentials and Cyber Essentials Plus certification regimes are all British products of the period.

In October 2016 CESG was reorganised into the National Cyber Security Centre (NCSC), the public-facing component of GCHQ headquartered at Nova South in Victoria. NCSC inherited the certified-information-assurance regime, the technical authority on cryptographic primitives in British government, and the relationship with CPNI (now NPSA) on critical national infrastructure. The British lineage of cryptographic trust now had a public face, a published research output, and a remit that explicitly extended to commercial and academic engagement. NCSC's CyberFirst programme, the Active Cyber Defence work, the published reports on Russia, China, Iran and the criminal threat landscape are all visible products of an institutional capability that has been continuous since 1940.

This matters to the Mickai story for a specific reason. The cryptographic primitives the Mickai substrate uses (FIPS 204 ML-DSA, hash-linked append-only ledgers, hardware-bound keys, attestation chains, browser-resident verifiers) are not novel inventions. They are the established post-quantum and post-attestation primitives that NCSC and its international peers have been preparing for. The novelty is the architectural composition of those primitives into a substrate that solves the audit problem for AI agents. Britain produced the lineage. Mickai composes the lineage into the substrate the AI era requires.

What changed in 2024 and 2025 that made the substrate necessary

Two specific things, beyond the obvious explosion of large-language-model deployments.

First, in August 2024 NIST finalised FIPS 204, the Module-Lattice-Based Digital Signature Standard, and FIPS 203, the Module-Lattice-Based Key-Encapsulation Mechanism. NCSC has been actively guiding UK organisations through the migration to post-quantum cryptography since 2019, with formal guidance updated in 2023 to recommend organisations begin planning the transition with a target horizon of the mid-2030s. The reason is straightforward. Cryptographically relevant quantum computing is expected to become available between 2030 and 2035. Any digital signature produced today under classical algorithms (RSA, ECDSA) will not survive that transition. Audit chains designed today need to be signed under the new standards, or they will become unverifiable in the operational lifetime of the systems they protect.

Second, between November 2023 and May 2026 the United Kingdom hosted two state-visit-grade AI safety summits, established the AI Safety Institute, funded ARIA's Safeguarded AI programme, and produced a sequence of policy documents that all assume the existence of a verifiable substrate underneath deployed AI. The substrate has not existed. The market has produced governance brochures and model cards. The lineage from Bletchley to NCSC, the institutional capability to treat trust primitives as something Britain makes, has been quietly waiting for a private actor to compose those primitives into the substrate the policy is asking for.

Workington, May 2026

The composition is now filed. The actor is Micky Irons (Mickarle Wagstaff-Irons), based in Workington on the Cumbrian coast. The substrate is thirty one UK patent applications at the IPO in Newport, totalling nine hundred and fourteen claims, sole inventor of record. Each filing carries a complete specification with description, claims, abstract, prior-art search, drawings, and Form-1 metadata. Each filing references the British lineage in its prior-art mapping. Each filing extends the lineage into a specific architectural composition that the AI era requires.

The headline composition is the audit substrate. Decision lineage with ML-DSA-signed causal audit ledger (GB2608804.7 / MWI-PA-2026-016) inherits the hash-linked append-only ledger pattern that British cryptographic engineering has been producing variants of since CESG's work on certified evidence systems in the 1990s. PQ-safe attestation and ML-DSA-signed tool-invocation ledger (GB2608806.2 / MWI-PA-2026-008) signs the chain under FIPS 204 ML-DSA-65, the standard NCSC's post-quantum guidance has been preparing British infrastructure for. Browser-resident offline post-quantum verifier (GB2610414.1 / MWI-PA-2026-023) takes the WebAssembly compilation pattern that Cambridge and Edinburgh academic groups have been advancing and applies it to the audit-verification problem, producing a verifier that runs in the procurement officer's browser without network access.

Trust-domain externalisation (GB2610415.8 / MWI-PA-2026-024) is the architectural commitment that the signing surface and the verification surface live outside the trust domain of the agent process that produced them. This is the same structural commitment that Bletchley operated under from the start. The codebreakers were in Hut 8. The verification of the breaks was in a separate, compartmentalised process. The intelligence consumers in London were in a third compartment. The substrate works because the compartments do not collapse into each other. Mickai's trust-domain externalisation is the modern form of the same commitment.

Voice-biometric-gated LLM tool invocation (GB2608799.9 / MWI-PA-2026-013) and per-skill clearance-gated execution (GB2608818.7 / MWI-PA-2026-021) extend the lineage in a direction the bombe-era engineers did not need but that the AI era requires. The actor invoking a high-impact action must be cryptographically attested at the moment of invocation, not at session start. The clearance regime that British government has operated on personnel for the entire post-war period is, in the Mickai composition, applied to AI tool invocations at the per-skill granularity. Trust does not cascade. Each invocation is gated against the current attested authority of the actor in the loop.

What this lineage commits Britain to do next

The lineage from Bletchley to Workington is not a heritage exhibit. It is an active institutional capability that, if Britain chooses to use it, can produce the global standard for AI audit substrate before the EU and the US do. The choice has three components.

**One, recognise the engineering layer as part of the policy.** AISI, ARIA, the AI Safety Summit secretariat, and the Cabinet Office have produced exceptional policy work. The next move is to recognise that the policy needs an engineering substrate to bind to, that the substrate is now in process at the IPO, and that the British institutional answer is to elevate the substrate into the procurement and evaluation methodology. None of this requires new legislation. It requires the institutional decision that the substrate counts as part of the policy work, not external to it.

**Two, fund the British engineers building the substrate.** Mickai is currently held privately by one British inventor of record with zero external investors. The lineage from Bletchley required Treasury funding through Cold War austerity. The lineage from CESG required Treasury funding through New Labour and Coalition austerity. The lineage from NCSC was funded by a 1.9 billion pound National Cyber Security Strategy commitment in 2016 and a 2.6 billion pound commitment in the 2022 strategy. The lineage from Mickai is, at the time of writing, funded by one inventor's personal capital. ARIA, the British Business Bank, the Future Fund Breakthrough programme, or a direct Treasury equity stake (as was used for OneWeb in 2020) are all available instruments. The substrate is a national strategic capability. The lineage suggests it should be treated as one.

**Three, lead the international standardisation.** The OAR schema is filed at the IPO. The browser-resident verifier is filed. The trust-domain externalisation pattern is filed. The path from a national filing to an international standard runs through the British Standards Institution, ETSI, ISO/IEC JTC 1 SC 27 (the working group on information security techniques), and the IETF. NCSC has long-standing positions in all of those bodies. AISI has emerging positions. The opportunity is to elevate OAR to a published BSI standard within twelve months and to seed the corresponding international work in the same window. The country that wrote the lineage gets to set the standard. Or it gets to wait while another country composes the same primitives and ships first.

Closing the loop

Eighty six years separate Hut 8 from Workington. The lineage is unbroken. The institutional capability that treated cryptographic primitives as a national strategic capability in 1940 is the same capability that produced the post-quantum signature standard British infrastructure is now migrating to. The composition of those primitives into a sovereign AI audit substrate is the next link in the chain. The link is filed. The next move is the institutional decision about whether the link gets recognised as part of the chain or whether it gets quietly absorbed into someone else's flag.

The lineage will continue either way. The question is which jurisdiction it continues in. The substrate is currently in process at the UK IPO. The country that produced Turing, Flowers, Cocks, Williamson and Ellis, that built Bletchley and Block H, that staffed CESG and NCSC, has the lineage to claim this work. The question is whether it does.

The inventor of record is contactable at press@mickai.co.uk. The portfolio is on mickai.co.uk/patents. The conversation is open.

“The lineage from Bletchley to Workington is not nostalgia. It is the only national tradition in the world that has continuously treated cryptographic trust as a thing the country builds. The AI substrate is the next link. It is filed. The decision about whether it gets recognised as British work, or absorbed into someone else's standard, is the institutional decision of the next six months.”

Sources and references

• Bletchley Park Trust, public archives on the bombe, Hut 8, and Alan Turing's role.
• The National Museum of Computing (TNMOC), Block H reconstruction of Colossus, Tommy Flowers archive.
• GCHQ, declassified history of public-key cryptography by Ellis, Cocks and Williamson, released 1997.
• NCSC, Migrating to post-quantum cryptography guidance, updated 2023 and 2024.
• FIPS 203, FIPS 204, FIPS 205, NIST post-quantum cryptography standards finalised August 2024.
• Cabinet Office, National Cyber Security Strategy 2016 to 2021 (1.9 billion pounds) and 2022 to 2030 (2.6 billion pounds) commitments.
• Mickai patent portfolio, mickai.co.uk/patents (31 filed UK patent applications, 914 claims, sole inventor of record Micky Irons).
• GB2608806.2 / MWI-PA-2026-008, PQ-Safe Attestation and ML-DSA Signed Tool-Invocation Ledger.
• GB2608804.7 / MWI-PA-2026-016, Decision Lineage with ML-DSA-Signed Causal Audit Ledger.
• GB2610414.1 / MWI-PA-2026-023, Browser-Resident Offline Post-Quantum Verifier.
• GB2610415.8 / MWI-PA-2026-024, Trust-Domain Externalisation Architectural Pattern.
• GB2608799.9 / MWI-PA-2026-013, Voice-Biometric-Gated LLM Tool Invocation.
• GB2608818.7 / MWI-PA-2026-021, Per-Skill Clearance-Gated Execution.
• ARIA, Safeguarded AI programme under chair Matt Clifford, programme director David Dalrymple.