Athens, Greece

Miltos Makridis

Broadcast journalism since 1987, and the technology underneath it for just as long. Today I work on AI governance — and I write the machine-readable artifacts that turn regulation into something a toolchain can execute.

Current work

Open standards · Apache-2.0

The EU AI Act, as a machine-readable catalog

An OSCAL catalog of the AI Act: 50 controls across Chapters I–V, IX and XII, five role-scoped profiles, and a readiness assessment plan. It encodes Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744, so a compliance toolchain can treat the Act the way it already treats NIST SP 800-53 or ISO 27001.

Permanent identifier w3id.org/eu-ai-act-oscal · source on GitHub

Worth knowing

The high-risk deadline moved sixteen months

The Digital Omnibus — Regulation (EU) 2026/1744, in force 27 July 2026 — rewrote Article 113(3). Annex III stand-alone high-risk systems now apply from 2 December 2027, not 2 August 2026. A good deal of published guidance predates that change and still gives the older date.

The corollary matters more: 2 August 2026 was not empty. Article 50 transparency landed then, and after the deferral it is the principal obligation that did. A readiness programme still working to the old high-risk date is working to a superseded one.

Not legal advice. The Official Journal governs; where the catalog and the Official Journal differ, the Official Journal is right.

Track record

Working with me

I take on fractional CTO and Chief AI Officer work, AI-governance and EU AI Act advisory, media × AI leadership, and scoped consulting engagements that end with something running.

Athens-based and an EU citizen, so no sponsorship is needed anywhere in the EU or EEA. Remote by preference, and open to travel or to relocating within the EU.