“The life so short, the craft so long to learn.” Geoffrey Chaucer wrote that more than 600 years ago, which is annoying, because it may be the best summary of the AI creative debate in 2026.
We keep talking about the tools as if access to them is the main story. Who has Runway? Who has Kling? Who has Leonardo.ai and Google OMNI? Who has the latest model, the better subscription, the faster workflow?
But craft has never worked like that.
Earlier this year, researchers from Deakin University and Swinburne University of Technology analysed 371 films submitted to Australia’s first AI Film Festival, DISRUPT.
Full disclosure: I helped create DISRUPT, so I have a horse in this race. But that also gave me a front-row seat to what the research later confirmed, and I share the learnings openly.
The headline finding wasn’t really about technology.
The best work came from the teams with the clearest creative purpose, not necessarily from the best tools. Everyone had access to roughly the same generation of AI tools: the Google stack, Runway, Kling, Leonardo.ai, ChatGPT, Midjourney, and the rest of the now-familiar creative AI stack. Several creators produced feature-length narrative films for under $1,000. The economics of creative production, already under pressure, were basically demolished in real time across 371 submissions.
And yet the work that stood out was immediately recognisable. Work that stood out because it had intention behind it. That wasn’t the magic prompt library and prompt engineering crap.
The research identified three human qualities that separated the strongest films from the technically impressive but forgettable ones: intention, taste and judgement.
That distinction matters for every brand and agency trying to work out what AI actually changes, how to hire, how to manage agency partners, and how to reorganise the operating model:
- Intention is clarity of creative purpose. It’s knowing what you’re trying to say before the machine starts saying things back at you. It means understanding both the potential and the limitations of AI, then using it.
- Taste is the harder one to fake. Taste is a human trait honed over years of practice, application, collaboration, and experience. It’s cultural capital. It’s the ability to read a category, understand its conventions, and subvert them with purpose.
- Judgement is the orchestration layer. Knowing which tools to use, when to use them, and when not to use them. Knowing when the first output is good enough to explore, but nowhere near good enough to ship. It’s restraint over spectacle.
This is where the AI conversation in marketing has become a bit lazy.
For the past two years, many organisations have treated AI creative capability as a production question. Can we make more content? Can we make it faster? Can we reduce cost? Can we version assets across channels without burning out half the studio?
Fair and possibly useful questions. But they’re floor questions, not ceiling questions.
If your AI strategy is built only around making more stuff, faster, congratulations. You’ve built a content treadmill. It may be efficient and even save money. But every competitor using the same tools can build the same treadmill.
The advantage isn’t access anymore. Access has flattened.
What hasn’t flattened is the quality and talent of the thinking happening before the prompt is written.
That has serious implications for how marketers hire. The instinct may be to look for younger, more technically fluent talent and assume they’ll naturally lead the AI creative shift. The DISRUPT data complicates that assumption. The average participant was 44 years old. The largest cohort was 40 to 49. Only 6 of 73 Australian submissions came from students.
That doesn’t mean younger talent isn’t critical. Of course it is. But it does suggest something important: AI amplifies the judgement people bring to it. If someone has spent years building taste, learning story, understanding culture, reading audiences, collaborating with clients and making hard creative calls that enforces and build distinctiveness, AI gives them more range. It doesn’t replace the accumulated craft.
That should change the hiring brief.
The next great AI creative hire may not be the person who knows every tool. It may be the person who knows what’s worth making and with core fundamentals.
It should also change how marketers engage their agency partners.
If the client brief becomes “use AI to make this cheaper and faster”, the agency will probably do exactly that. But the output will sit in the great grey middle of competent AI-assisted work. Faster, cheaper and forgettable is procurement wearing a hoodie.
The better question for agencies is different.
Where does AI genuinely expand the idea? Where does it let us prototype faster, explore stranger territories, push the limits of differentiation and distinctiveness, build richer worlds, or test more directions without settling too early? And where should we deliberately slow down because the work needs human friction, debate and taste?
Most bad AI work exists because the human accepted the first acceptable answer. The machine produced something OK, and everyone quietly agreed to stop there. It’s like ordering a banquet meal and leaving the restaurant after the bread basket.
This is where operating models need to change.
Most marketing teams are still structured around old production scarcity. Briefs move slowly, with approvals piling up and creative treated as a sequence of handovers rather than a live system of judgement. Then AI gets dropped into the middle and everyone wonders why the work is faster but not necessarily better.
There will be moments where AI provides a great solution for genuine needs, whether it’s creative at scale, retouching, ecommerce SKU placements, or operational propositions.
And that is the friction between convention and innovation, the abundance of doing more and cheaper versus knowing what not to make.
The full DISRUPT research is downloadable. But the short version is this: 371 films, similar tools, similar access, and the work that mattered still came down to whether a human being had something worth doing.
AI hasn’t made creative judgement less important. It’s made the absence of it much harder to hide.
Lucio Ribeiro is Chief AI & Innovation Officer at TBWA\Australia and lecturer at RMIT.
