# AI regulation, standards and reality Bridging the gap between regulatory frameworks and practical implementation. Exploring how standards, assurance, and compliance actually work in the real world - from AI to cement. (Active AIQI members drop me an email for complimentary upgrade) Page: Feed: --- ## [ISO/IEC 4213 2nd edition reaches Draft International Standard stage](https://adamleonsmith.substack.com/p/isoiec-4213-2nd-edition-reaches-draft) _2026-08-03 · Adam Leon Smith DEng FBCS_ From a technical specification to the standard nobody can avoid ## [prEN 18229-3 reaches Enquiry](https://adamleonsmith.substack.com/p/pren-18229-3-reaches-enquiry) _2026-07-30 · Adam Leon Smith DEng FBCS_ The standard that admits human oversight sometimes cannot work ## [Revised AI Act text available](https://adamleonsmith.substack.com/p/revised-ai-act-text-available) _2026-07-24 · Adam Leon Smith DEng FBCS_ My shortest article of all time. ## [EN 18286:2026 is finally published, the first AI Act standard to reach that milestone](https://adamleonsmith.substack.com/p/en-182862026-is-finally-published) _2026-07-22 · Adam Leon Smith DEng FBCS_ What changed since the public enquiry version? ## [prEN ISO/IEC DIS 23282: How to measure whether Natural Language AI is accurate](https://adamleonsmith.substack.com/p/pren-isoiec-dis-23282-how-to-measure) _2026-06-30 · Adam Leon Smith DEng FBCS_ What the new NLP evaluation standard covers, and why providers of high-risk and general-purpose AI should read it ## [Two standards, one architecture: FprEN ISO/IEC FDIS 24970 and prEN 18229-1](https://adamleonsmith.substack.com/p/two-standards-one-architecture-fpren) _2026-06-05 · Adam Leon Smith DEng FBCS_ Two significant milestones landed this spring serving both the state of the art and Article 12 of the EU AI Act. ## [9 things to look for in compliant agentic AI](https://adamleonsmith.substack.com/p/9-things-to-look-for-in-compliant) _2026-06-02 · Adam Leon Smith DEng FBCS_ This article builds on the paper "AI Agents under EU Law" and a keynote presentation I gave to market surveillance authorities in May ## [You can't do AI ethics without AI ethicists](https://adamleonsmith.substack.com/p/you-cant-do-ai-ethics-without-ai) _2026-06-01 · Enrico PANAI_ EN 18274 approaches publication ## [The Commission reviewed its AI list. Therapy chatbots are the problem it couldn’t classify.](https://adamleonsmith.substack.com/p/the-commission-reviewed-its-prohibited) _2026-05-29 · Adam Leon Smith DEng FBCS_ Self-help therapy chatbots are flagged for monitoring. ## [First EU AI Act cases in front of the courts](https://adamleonsmith.substack.com/p/first-eu-ai-act-cases-in-front-of) _2026-05-25 · Adam Leon Smith DEng FBCS_ C-806/24 and C-245/25 are the first cases at the CJEU. Here’s what the judges will be deciding. ## [Transcript-Driven Stock Discovery (Sponsored)](https://crawlproof.com/a/8jgCcB60sIFC) _2026-05-25 · **Sponsored**_ Ranked, evidence-backed watchlists from indexed executive transcripts for 1–2 quarter horizons ## [A general intended purpose includes all high risk use cases by default](https://adamleonsmith.substack.com/p/a-general-intended-purpose-includes) _2026-05-21 · Adam Leon Smith DEng FBCS_ The Commission’s guidance on high-risk classification dropped consequential clarification on intended purpose. ## [James Gealy On AI Safety Standards For Frontier Models](https://adamleonsmith.substack.com/p/james-gealy-on-ai-safety-standards-2e7) _2026-05-20 · Adam Leon Smith DEng FBCS_ In this episode, hosts Michael Mainelli and Adam Leon Smith welcome James Gealy, Standardization Lead at Safer AI, a French NGO focused on AI risk modelling. [Listen](https://api.substack.com/feed/podcast/198592137/0013549d75086f9a2f6a978b84326a52.mp3) ## [The Standard that watches the watcher](https://adamleonsmith.substack.com/p/the-standard-that-watches-the-watcher) _2026-05-20 · Enrico PANAI_ ISO/IEC DIS 25029 and the governance of AI-enhanced nudging ## [The pace is picking up - JTC 21 Timeline Update – May 2026](https://adamleonsmith.substack.com/p/the-pace-is-picking-up-jtc-21-timeline) _2026-05-15 · Adam Leon Smith DEng FBCS_ More than half of the standards responding to the EU AI Act are now in public enquiry, or later in the process. ## [prEN 18282 heads to Enquiry: Cybersecurity specifications for AI Systems](https://adamleonsmith.substack.com/p/pren-18282-heads-to-enquiry-cybersecurity) _2026-05-08 · Adam Leon Smith DEng FBCS_ The latest harmonised standard to go to public enquiry supports compliance with Article 15 ## [What was actually agreed in the Omnibus and what does it mean?](https://adamleonsmith.substack.com/p/what-was-actually-agreed-in-the-omnibus) _2026-05-07 · Adam Leon Smith DEng FBCS_ Analysis of the full leaked compromise ## [prEN 18228 heads to Enquiry: the product-safety answer to AI risk management](https://adamleonsmith.substack.com/p/pren-18228-heads-to-enquiry-the-product) _2026-05-07 · Adam Leon Smith DEng FBCS_ The draft harmonised standard on AI risk management has reached the CEN Enquiry stage. Here’s what practitioners should notice about how it defines risk, what counts as a risk control, and what to do ## [Mike Thieme On Making AI Standards Work For Busy Professionals](https://adamleonsmith.substack.com/p/mike-thieme-on-making-ai-standards-1a5) _2026-05-06 · Adam Leon Smith DEng FBCS_ In this episode, hosts Michael Mainelli and Adam Leon Smith welcome Mike Thieme, Managing Director and Cloud Advisory Lead at Accenture. [Listen](https://api.substack.com/feed/podcast/198592138/00e20770ea14f64d911b192438a771a4.mp3) ## [ISO/IEC 25059 gets a rewrite: AI quality models expand beyond the product](https://adamleonsmith.substack.com/p/isoiec-25059-gets-a-rewrite-ai-quality) _2026-05-04 · Adam Leon Smith DEng FBCS_ The SQuaRE quality model for AI systems revision completes enquiry ## [Refer a Friend, Earn Rewards (Sponsored)](https://crawlproof.com/a/uyzWS4FwaPP4) _2026-05-04 · **Sponsored**_ Earn rewards when friends get the Graphite Business Card through your referral. ## [AI Agents Under EU Law: what providers actually have to do](https://adamleonsmith.substack.com/p/ai-agents-under-eu-law-what-providers) _2026-05-02 · Adam Leon Smith DEng FBCS_ High-risk agentic systems with untraceable behavioural drift cannot currently satisfy the essential requirements of the AI Act