For the last two years, AI literacy has been taught as a technical skill. Learn the tools, master the prompts, keep pace with whatever platform launched last week. We think this framing has quietly become the wrong one, and it is starting to cost agencies and brands real credibility with the people they most need to reach.
The technical side of AI is becoming commoditised at speed. What is not becoming commoditised, and what is in fact becoming rarer and more valuable, is the ability to translate what a technology does into what it means for the person using it. That translation is not a soft skill sitting politely beside the strategic work. In our experience it is the strategic work, and it is where leadership in this industry will increasingly be won or lost.
We have spent our career closing that particular gap, and we have come to call it Humanity Bytes. It is a simple enough idea. As platforms and algorithms grow more sophisticated and more invisible, the human meaning behind them has to work harder to stay legible. Brands that treat this as a communications afterthought, something to be handled once the technology is built, tend to struggle. Brands that treat it as core strategic infrastructure tend to earn trust faster and keep it longer.
Nowhere was this more apparent to us than in our work with HarmBlock, the ethically trained algorithm developed by SafeToNet Labs, which can be embedded into any operating system to detect and prevent the sharing of child sexual abuse material before it happens. We are proud that this work was recognised with a Creativepool award, and we mention that not to boast but because the award only confirms what the brief itself demanded from the start.
The technology behind HarmBlock is genuinely complex. It involves machine learning models trained on ethical frameworks, real time detection, and integration at the operating system level across an enormous range of devices and platforms.
None of that complexity is something a parent, a safeguarding lead, or a device manufacturer needs to carry in their head. What they need to understand, instantly and without translation fatigue, is that a piece of technology now exists that can stop harm before it happens, wherever it is embedded.
Our job was to hold the full technical weight of that innovation on one side, and to deliver something clear, human and urgent on the other. That is the whole discipline in a single brief. Get the translation wrong and a genuinely life saving technology sits unused because nobody outside the engineering team can explain why it matters. Get it right and the technology travels far beyond the room it was built in.
We saw the same challenge from a different angle in our work with Suslop, a Canadian sustainable development consultancy working on major infrastructure such as small modular reactors and subway extensions. Infrastructure of this scale is, by nature, mostly invisible to the public until it goes wrong.
Our response was a concept we called The Invisible Made Visible, built around a fingerprint encoded mark that made the human fingerprint on large scale, long term infrastructure decisions felt rather than assumed. Once again, the brief was not really about the engineering. It was about making sure the people affected by that engineering could feel the intention and care behind it.
Our work with Seyi Akiwowo, founder of Glitch and now a governance advisor working across boards and institutions, sits in the same territory, even though the subject matter is governance rather than a product. The brand essence we developed, Rebuilding with Care, treats care itself as infrastructure. Good governance, like good technology, is often complex, procedural and hard for outsiders to parse. The leadership skill here is the same one required of HarmBlock and Suslop. Take something structurally complicated and make its human intention unmistakable, without flattening the substance that makes it trustworthy in the first place.
This is why we think AI literacy, properly understood, belongs at leadership level rather than being delegated to whoever is closest to the tools. A creative director, a founder, or a brand lead who can hold both the technical fluency and the human fluency at once is going to make better calls about what to build, what to say, and what to leave out. Agencies who can offer that dual fluency will increasingly be the ones trusted with the work that matters most, particularly for organisations building technology for good, where the stakes of getting the translation wrong are measured in trust lost rather than clicks missed.
The tools will keep changing. The organisations we most want to work with, the ones building ethically trained algorithms, sustainable infrastructure and better governance, will keep needing someone who can stand between the engineering and the public and make the human story impossible to miss. We think that is the leadership skill worth building for 2026, and it is the one we intend to keep practising.