A technical contribution for standards processes.
aiGUARD Systems is committed to active engagement with the standards bodies whose specifications define how AI governance is implemented in regulated industries.
Why standards engagement is core to our strategy.
The EU AI Act, the FCA Consumer Duty, and the Public Sector Bodies Accessibility Regulations all create legal obligations for which no standardised technical solution currently exists.
Regulators and industry bodies do not solve this gap by inventing standards from scratch. They convene working groups, evaluate available technical contributions, and ratify those that meet the relevant criteria. For the gap between Article 12 of the EU AI Act, which requires automated logging of high-risk AI activity, and the available standardised solutions, we have not identified a published standard that provides per-inference, externally verifiable governance certification in the form proposed in our working paper.
aiGUARD Systems Limited intends to be a recognised technical contributor to the standards that close this gap. Our patent-pending architectures, Thames Sentinel™ (aiGUARD™), the Governance Execution Certificate (aiGEC™), and Synapse-ID™, were designed from the outset to be standardisable: open at the interface, deterministic in behaviour, and verifiable without access to the issuing system.
Where the work happens.
Standards-body engagement is iterative, technical, and consensus-driven. The following bodies and committees represent our priority targets for technical contribution:
BSI ART/1, Artificial Intelligence
The UK national mirror committee for artificial intelligence, feeding into ISO/IEC JTC 1/SC 42. Responsible for UK national input into international AI standards. Priority engagement target for aiGUARD™ and aiGEC™ contributions on AI output governance, logging, and certification.
ISO/IEC JTC 1/SC 42, Artificial Intelligence
The international joint technical committee for AI standards. The committee's published work includes ISO/IEC 42001 (AI management systems) and ISO/IEC 23894 (AI risk management), alongside ongoing work on AI transparency, governance, and related topics. Strategic destination for international standardisation of the trust stack.
ISO/IEC JTC 1/SC 27, Information security, cybersecurity and privacy protection
The international committee for information security and privacy. Relevant to Synapse-ID™ technologies for identity footprint reduction and architectural privacy. Engagement target for Synapse-ID™ contributions on architectural privacy and identity-by-design.
Mastercard / Google Verifiable Intent
An open industry specification for agentic commerce trust, published February 2026 under Apache 2.0. The specification defines an optional agent_attestation field intended to carry agent identity or security attestations, content currently deferred to future companion documents. aiGEC™ is a candidate technical contribution to populate the agent_attestation field.
UK government & sector bodies
The UK AI Standards Hub (BSI / National Physical Laboratory / Alan Turing Institute), the FCA on financial services AI, NHS England on healthcare AI, and the Central Digital and Data Office on public sector accessibility. Engagement on sector-specific application of the trust stack standards.
Fair, reasonable, and non-discriminatory licensing on standards-essential patent claims.
aiGUARD Systems Limited makes the following commitment in respect of any patent claim that becomes essential to a published standard:
Where any claim of a patent owned or controlled by aiGUARD Systems Limited is determined to be essential to a published standard issued by BSI, ISO, IEC, ITU, ETSI, or an equivalent recognised standards body, aiGUARD Systems Limited undertakes to license such claim on Fair, Reasonable, and Non-Discriminatory (FRAND) terms to any party implementing the standard, in accordance with the policies of the relevant standards body.
This commitment applies only to patent claims that are essential to a relevant standard, that is, claims whose practice is required to implement the standard. It does not apply to patent claims that read on aiGUARD Systems products, deployments, or implementations outside the boundary of an essential standard claim.
Commercial licensing of the trust stack products and their non-standards-essential patent claims is governed by our Licensing policy. The two regimes operate in parallel: FRAND for standards-essential claims; commercial licensing for everything else.
aiGEC™ into agent_attestation.
Our first concrete standards contribution is our working paper AGS-PP-001 v1.0: Closing the EU AI Act Article 12 Logging Gap, published on Zenodo (DOI 10.5281/zenodo.20140622) under CC BY 4.0 on 12 May 2026. The paper articulates four architectural properties as the minimum required for Article 12-compliant per-inference logging and presents per-inference governance certification (aiGEC™) as a candidate technical specification.
Following on from the paper, our next concrete contribution targets the agent_attestation field defined in the Mastercard / Google Verifiable Intent specification. The specification defines the field as an extension hook for carrying agent identity or security attestations, with content deferred to future companion documents.
aiGEC™ is presented as a fully specified candidate contribution, technically aligned with the goals of Verifiable Intent. A Verifiable Intent credential carrying an aiGEC™ attestation would provide the receiving payment network with cryptographic proof that the AI output generating the transaction instruction was evaluated through a non-bypassable governance pipeline before delivery.
We are preparing a technical submission to the Verifiable Intent specification process documenting the aiGEC™ specification, its applicability to the agent_attestation extension, and the FRAND licensing terms under which the relevant patent claims would be made available.
Standards engagement timeline.
| Period | Activity | Target outcome |
|---|---|---|
| Q2 2026 | BSI membership | Application for BSI membership and request for ART/1 participation |
| Q2–Q3 2026 | Verifiable Intent submission | Technical contribution to the agent_attestation working group, proposing aiGEC™ as reference content |
| Q3 2026 | Public consultation responses | Formal responses to relevant BSI, ISO, and EU AI Office calls for evidence on AI governance, logging, and transparency standards |
| Q4 2026 | PCT international filing | PCT entry for the GB2603184.9 anchor patent, preserving rights for national-phase entry aligned with target standards bodies |
| 2027 | New Work Item Proposals | Preparation of candidate New Work Item Proposal materials for submission through the appropriate BSI/ISO route |
| 2027–2028 | Working draft contributions | Technical drafting contributions through Working Draft, Committee Draft, and Draft International Standard stages |
Standards engagement timelines depend on committee schedules, working group consensus, and the pace of regulatory development. The roadmap above represents our current planning assumptions and will be revised as work progresses.
Engage with our standards work.
Standards bodies, working group secretariats, regulators, and industry partners interested in aiGUARD Systems’ standards contributions are invited to make contact.
Contact us