While AI regulation addresses fake things that seem real, my challenge is that real things can look fake.

As AI-generated images become more realistic, laws such as the EU AI Act are introducing disclosure requirements intended to make it clear when content has been generated or manipulated by AI.

But I’ve got the opposite problem.

You don’t need to have followed me for long to know that I love using images of minifigures from my favourite toy to illustrate my posts. In the spirit of 'what you see here is what you get when you hire me', they’re also a key feature of my keynotes.  Nobody thinks they’re real. 

But my keynotes and workshops don’t just use minifigures. Some of my most memorable and impactful slides feature Compliance In The Wild; ridiculous signs, absurd situations and spectacular examples of human behaviour. Things that are 100% genuine, but seem so implausible that I’ve always felt I had to reassure audiences they aren’t fake. It’s their very implausibility that makes them so powerful.

Recently, my reassurance has shifted from 'I promise this is real' to 'this is not AI-generated.' Rather than saying it every time and disrupting my flow, I’m creating an icon to do it for me. An AI-generated icon, obviously.

It’s also an interesting Compliance experiment: to test whether, for my audience and the materials I use, deploying the opposite logic to the rules is actually more useful.