Much of our published content now features a 'human-made' badge. This page explains what we mean by that.
AI slop is everywhere now, and a lot of it is wrong. So, going forwards, we're labelling our content for what it's always been: human-made, by people on our team.
For a small team like ours, a cloned voice or AI-generated visuals are the difference between publishing twice a month and twice a week. The temptation to cut corners is very real.
But at our core, we've always been about providing the human touch, and we think that has to extend to the things we make, not just how you interact with us.
Our videos are all scripted and storyboarded by hand, with audio recorded by Charles, and audio processing, illustration, animation and final video editing all done by Adam.
It's worth noting however, that we do use AI tools for small, 'de minimis' use in this process, to reduce time on tasks that would have little-to-no impact on the authorship of the finished work.
De minimis is from the Latin maxim de minimis non curat lex, meaning "the law does not concern itself with trifles". For us this isn't about the law though, it's about being open and honest regarding our limited use of AI.
This page exists to clarify that usage policy.
Some definitions of human-made hold themselves to a 90%/10% ratio, where AI tools are used in the process, but it's mostly human generated.
Our own usage is closer to 95%/5%. We see AI as a tool to assist with content we've made ourselves, rather than a content generating machine.
We may use AI tools selectively for internal functions in the business, to improve productivity. But to be clear: we will never we never use customer data to train AI, and we don't put customer data into AI tools.
Some of the third-party hardware and software we provide to clients may have AI tools built in, such as AI malware detection. We can't strip those out, and in most cases wouldn't want to. If you'd like to know more about what's built into the products we use, ask and we'll let you know.
If something we've published looks AI-made to you, let us know. We'd rather hear about it and put it right than have you wondering.
Version 1.0
Updated 19th August 2026
Any significant changes will be noted here.
