- AI Competence Measurement Standard for EU AI Act Article 4 Compliance — cont. to CEN-CENELEC JTC 21
Artificial Intelligence
AI literacy and competent human oversight are legal duties under the AI Act (Articles 4, 14 and 26(2)), but there is no standardised way to measure whether the people exercising oversight actually hold the competence the law assumes. This fellowship names that gap and offers an open, candidate basis for closing it.
The benefit to the EU, and especially to SMEs, is proportionality and portability. A risk-based instrument concentrates expensive assessment only on oversight-critical roles and clears the long tail by attestation and sampling, so a small firm can evidence Article 4 "best-efforts" diligence without assessing every employee. Because it is published CC BY-SA 4.0, there is no vendor lock-in; competence can be compared across roles and organisations on a common protocol; and an organisation gains a defensible diligence record for a market-surveillance authority — more than a training list. It lowers the barrier to adopting AI responsibly.
- New standard development
-
CEN-CENELEC JTC 21 (Artificial Intelligence), reached through Latvia's national mirror committee LVS/STK 59, which I joined in July 2026. Affected drafts: prEN 18228 (AI Risk Management, AI Act Art. 9) — comments lodged on 4.3 "Competence of personnel" and on post-market monitoring, with proposed normative text, now carried in Latvia's national position; and prEN 18229-1 (Logging, transparency and human oversight, Arts 12-14), enquiry open, comments in preparation. A study work item on competence measurement is under discussion at LVS.
Title & Organisation Name: Idea Jet Lab SIA
Country: Latvia
