Field notes from inside a sovereign AI.
Mickai® is a Sovereign Intelligence Operating System (SIOS) that runs on your own hardware. These are long-form essays on the architecture of the Mickai SIOS, the patterns that keep surfacing in commercial AI in 2026, and the engineering choices that make sovereign intelligence possible. Written by Micky Irons, named inventor of the 104 filed UK patent applications, recorded on the UK IPO public register at numbers GB2607309.8 to GB2611702.8, GB2611885.1 onwards, GB2612762.1 to GB2612793.6, GB2613386.8 to GB2613404.9, and GB2615041.7 to GB2615043.3.
The cooperative the field notes describe, running on the sovereign silicon substrate.
Showing 241 to 264 of 1124 articles.
Your data becomes your advantage when it stays in one operating system
When all your company data lives in one owned operating system, the assistant gets sharper and the business gets better informed. Here is why.
AI Disaster Recovery and Business Continuity: The Complete Guide
AI dependent processes need the same BIA, RTO and RPO discipline as any other critical process, plus one metric classic DR never had to define.
AI Governance Framework for Enterprise
A named owner per system, an applied risk tiering rule and an audit ready evidence trail turn a policy document into an operating model.
AI Procurement Checklist (2026): The Complete Buyer's Framework
Five questions, worked through data, deployment, ownership, audit, security and exit terms, turn AI vendor evaluation into a defensible process.
AI Risk Register Template (Free Download)
Five columns, a 5x5 scoring matrix and a named review cadence turn a spreadsheet into a working AI risk register.
AI Security Checklist for CISOs
A six layer attack surface map and its controls turn a board mandate for AI security into a working programme.
Are AI embeddings personal data?
Very often yes, and the safe default is to treat embeddings as personal data unless proven otherwise.
Build vs Buy vs Rent AI: The Enterprise Decision
Five variables, not a single rule, decide whether to build, buy or rent AI, and most organisations need more than one answer.
Can AI records be used as evidence in court?
Yes in principle in England and Wales, but the real battle is authenticity and weight, not admissibility.
Can barristers use AI on case papers and instructions?
Yes for drafting and organising, provided privileged papers never leave a boundary the barrister or chambers controls.
Can councils use AI to process planning applications?
Yes for summarising, drafting and checking inside the council's own boundary, and no for the decision, which must remain visibly human.
Can dental practices use AI on patient records?
Yes for note drafting and letters, provided special category health data never reaches a consumer cloud tool.
Can family law teams use AI on children and financial remedy cases?
Yes for drafting and chronologies, but only inside the firm's own boundary given how confidential these files are.
Can GP practices use AI scribes for consultations?
Yes, provided the practice acts as the data controller it is: DPIA first, patients told, and every note verified by the GP.
Can veterinary practices use AI on clinical records?
Yes, and it is a clean example of sovereignty as good practice rather than a compliance requirement.
Can you answer a subject access request about AI processing?
Only if you can search everywhere your AI puts personal data, and most organisations cannot yet do that.
Can you stop staff using AI on personal devices?
No, not completely, and the strategy that works out-competes the personal phone rather than trying to police it.
China now regulates AI agents. What should UK boards take from it?
China's new agent rules point to a shape UK boards should adopt now: a mandate, an identity and a reviewable record for every agent.
The Complete CIO's Guide to On-Premise AI
Everything a CIO must decide and defend to run AI on infrastructure the organisation owns, from GPU sizing to the audit trail.
Does cyber insurance cover AI incidents?
It depends on wording written for human-triggered incidents, and silent AI risk is becoming the new silent cyber.
The Enterprise AI Maturity Model
Five levels from isolated pilots to independently verifiable, owned infrastructure, with the unlock test for each.
The EU AI Act Explained for CEOs (2026)
A binding regulation applying in stages, not on one date, with high risk obligations now due in 2027 and 2028.
The Future of AI Infrastructure, 2026 to 2035
Infrastructure control migrates from vendor to operator across the decade, driven by regulation, not preference.
How do you acceptance test an on-premise AI before go-live?
Run the go-live gate, eight checks with pass criteria fixed before testing starts, or it was a demo, not a test.