Why your intern’s AI images aren’t up to standard

If you were hoping for a -yet another guide on how to improve your AI visual prompts- sorry to disappoint, but this won’t be it.
More of an ode to visual practice and eye training in a landscape where “use this reference and replace the product for this” prompts are expected to define a campaign’s visual identity.

I like to think of creative brains as a 1940s ENIAC; constantly processing information, testing pairings and the unpairable, imagining the unimaginable, making something you ate when you were 3 somehow relevant to something you heard when you were 30. And by a creative brain I don’t mean someone’s who went to uni and proved they knew by heart all of -insert here Western country established as the go-to in art history- most famous artists from the 60s. It may simply be someone with the ability to see the world beyond surface level.

It appears with the surge of AI and new visual generative technologies, the stronger the uncertainty about the creative process is growing. Everyone can now generate something that ‘looks like a campaign visual’. I’d like to disagree. Just the fact that we’ve democratised access to “bespoke” image generation, doesn’t necessarily mean everyone with access to it has the visual language or reference archive to even be able to articulate prompts, and ultimately create worth-using visuals.

| Reference Archive (noun)
A creative’s internal library of references, built through years of curiosity rather than collected through searches. It is less about remembering images than remembering why they mattered—how they felt, what they communicated and the unexpected connections they continue to inspire.

The longer I work in the industry the louder I hear Miuccia Prada’s words in my head “Study, learn, watch movies, art and read literature”. Collecting references, nurturing curiosity and refreshing inspiration is so important for creative brains, it is a way of not only seeing things from different perspectives and building new associations, but also to create and build your reference archive. And it’s time consuming, but so is any investment.

Probably that is why I struggle when I hear “we’ll just get the intern to do it in AI”.

Not because junior creatives don’t know how to use the tool, on the contrary. But expecting someone at the very beginning of their creative career to suddenly articulate and produce work with the depth, nuance and visual language of someone with over +15 years spent training their eye, is greatly misunderstanding what creativity actually is.

We suddenly have a tool capable of imagining the unimaginable, yet so much of its potential is spent recreating images that already exist. Originality has never come from the tool itself, but from the person using it.

Creative freedom, experience winning and failing, and time are what build your visual archive to eventually differentiate between “why a visual stays with people” vs “how can you recreate an existing image”.

If you’re not hiring juniors, you’re missing the great opportunity of having the next generation exploring, experimenting and developing new ways of seeing. To build their reference archive that will one day become unique and unexpected ideas.

AI can generate visuals in seconds, but what it can’t generate is new ideas after years of observation, associations and visual memory that tell you why a visual should exist in the first place.

Perhaps that’s the opportunity hidden in all this AI landscape mess. If AI increasingly generates more visuals, then perhaps our value as creatives shifts to judgement, curiosity, cultural literacy and our ability to connect ideas that have never been done before.

The future of creativity may not be about protecting the process from new tools, but about protecting the people who make the process meaningful and unique in the first place.