Blog

Global AI regulation will not stay fragmented. This paper forecasts its coalescence around three regulatory blocs on staggered timelines, and the bipolar interregnum that precedes the full structure.

Ungoverned AI is now a top-tier institutional risk. This paper sizes the exposure across thirteen categories and shows where governance investment produces the largest reduction in expected loss.

A single AI system rarely sits under a single rulebook. This paper sets out TRACE, a five-phase methodology for governing AI across overlapping frameworks, and a maturity model for locating where an organisation stands.

Agentic AI reshapes work, not just tools. This paper offers a governance and workforce framework: principles, board fluency, role and career architecture, and a maturity pathway.

Recruitment and workforce-management AI falls within Annex III. This paper maps what is in scope, how the Act meets GDPR and equal-treatment law, and the data-governance duties that follow.

Credit and insurance AI sits squarely in Annex III. This paper maps what falls in scope, how the Act meets EBA and EIOPA guidance, and how to manage a large model portfolio.

Article 10 and the GDPR pull in opposite directions on training data. This paper maps the conflicts, including representativeness versus minimisation, and offers a practical framework.

The AI Act is enforced through layered oversight: national authorities, the AI Office, and the AI Board. This paper maps how supervision will work and what draws regulatory attention.

Most EU AI Act guidance targets providers. Article 26 places eight complex obligations, plus the Fundamental Rights Impact Assessment, on deployers of high-risk AI. Here is where to begin.