This newsletter provides practical guidance, tools and resources for the real work of governing safe, secure and lawful AI.

Newsletter #55 - April 2026

Feature Article #1

The biggest risk in AI governance is waiting for the perfect moment to start.

By James Kavanagh

There have never been more frameworks, more conferences, or more confident agreement that AI governance matters. And yet, it's not clear that the talk is turning into action.

I've been in those meeting rooms and conference halls. The panels where smart people debate the merits of AI regulation, present responsible AI principles, and nod in unison about urgency. And then everyone flies home and does nothing different. Meanwhile, dev teams are shipping AI systems every week. Agents are being pulled into customer-facing products. Tools that risk and compliance have never seen are quietly embedded in how people work.

The gap between the conversation about AI governance and the practice of it has never been wider. This article is about why - and what to do about it.

I identify three traps that keep organisations frozen: waiting for conditions that will never arrive, treating governance like a project with a completion date, and convincing themselves they need more expertise before they can start. I explain why all three are mistaken ideas, and how organisations can start building real governance capacity right now - without waiting for the perfect moment of certainty and preparation. That moment isn't coming.

Feature Article #2

Building AI governance while building AI.

By James Kavanagh

In 1932, eleven ironworkers ate lunch on a bare steel beam 850 feet above Manhattan. No harnesses. No nets. No guardrails. People look at that photograph now and see recklessness. I think it misreads the history.

Those men weren't ignoring safety. Safety, as we understand it today, simply didn't exist yet. It couldn't be invented in the abstract, in an office, by people who'd never been up there. It had to be built alongside the work, by the people doing the work, as they learned what the work actually demanded. The formal codes came later. But governance - local, practical, evolving - was already happening, in the practices that kept people alive.

That's the history that matters for anyone building AI governance today. Not because the analogy is perfect, but because it answers the most common excuse for inaction: we don't know enough yet to govern properly. The ironworkers didn't know enough either. They built anyway.

This article is builds on my previous one about not waiting for the perfect moment. It's about the sequence of construction - what to build first, how to make it functional, and how to design governance that learns and improves rather than calcifying into another layer of bureaucracy. I go through how to map what AI you actually have, how to build practitioner capability alongside organisational capacity, and how to mechanisms every AI governance function needs, with feedback loops from day one.

From the Blog

Prepare for and pass the IAPP AIGP exam in 21 days.

By James Kavanagh

The IAPP AIGP Certification whether you have foundational knowledge across four areas: what AI systems are and how they're used responsibly, the basics of how laws and frameworks apply to AI, how governance fits across the AI lifecycle. It's a useful starting point of knowledge, that you can build on as a practitioner building real capability and skill.

But most AIGP candidates spend far longer preparing for this exam than they need to. And the reason is usually the same: their prep materials spend too much time on Domain I - foundational concepts, terminology, context - which accounts for only around 17 to 21% of the marks. Domains III and IV, which together carry more than half the exam, get underweighted. Plus, some candidates spend most of their study time watching videos and reading reference sources that are unexamined, rather than practising scenario questions, which is where the exam is actually won or lost.

This article is my practical guide to passing the IAPP AIGP certification in 21 days using the AI Career Pro Exam Prep course. Thirteen study days, two rest days, five full practice exams. A day-by-day breakdown of what to cover, where to invest more time, and how to use the practice exams to identify and close gaps rather than just measure them. It's built around the reality of how the exam is structured and weighted - not around a course design that treats Domain I as the main event.

If you're considering the AIGP certification, or you're mid-preparation and wondering why it feels harder than it should, this is worth reading before you spend another week studying the wrong things.

​Read the full article​

What we're working on

The AIGP Exam Preparation course is now fully live. All five practice exams are available including one that's free, and after a few weeks of calibration, the average score sits at 74% now. That's not an accident. A practice exam that everyone passes first time isn't doing its job. If you're considering the AIGP Certification, try the full 100-question free practice exam now and see where you actually stand before you sit the real thing.

The thing I mentioned in our last edition - something to help practitioners and organisations understand their readiness and where to focus - is nearly ready. I'll have more for you next week. I'm sure you'll like it. Standby.

As always, if anything in this newsletter or in any of the work you're doing with us has raised questions or sparked something in you, please reply to this email. We read everything and are always glad to hear from the community.

PS. You're receiving this as a subscriber to communications from AI Career Pro. We respect your privacy, so please unsubscribe through the link below if you do not wish to receive these communications in the future.