Guides
Plain-English answers to the questions a compliance owner actually searches: how the frameworks compare, what a duty really requires, and what defensible evidence looks like. Every fact is dated and cited.
The uneven transformation: the case for institutional adaptation
Neither utopia nor apocalypse: a technology as important as electricity, deployed as unevenly as the internet, governed as clumsily as everything else. The four contingencies worth fighting over, and the rational posture. The GAGE position.
How to document AI literacy for EU AI Act Article 4
Article 4 is proven with records, not a badge. The six-part compliance record, what counts as evidence, and where a verifiable credential fits. Sourced and dated.
EU AI Act vs NIST AI RMF vs ISO 42001
Binding law, a voluntary framework, and a certifiable standard. What each is, who it binds, and how they stack, with sources and dates.
AI literacy policy template (Article 4 ready)
The eight sections a defensible AI literacy policy needs, what each covers, and how to evidence the training parts. Copy the outline and adapt it.
What is a verifiable AI literacy credential?
A record an outside party can independently confirm and inspect: tamper-evident, resolvable, and showing the assessments passed. Why it beats a certificate of attendance.
Does the EU AI Act require AI literacy training?
Yes, Article 4: who is covered, what 'sufficient' literacy means, the deadline, and why training is no automatic presumption of compliance. Answer-first and sourced.
Agent skill governance: how to control what your agents install
Skills run with the agent's full privileges and publish with almost no review. The six controls a defensible skill lifecycle needs, from inventory to retraction, with the OWASP and Snyk evidence.
AI skills assessment: what to measure, and how
The five dimensions a real AI skills assessment measures, how it differs from a completion quiz, and how to get a scored result you can put on record.