The Economic Burden Hiding Inside AI-Only Work

I started using artificial intelligence in my own work, just like everyone else, and I'll admit upfront that I don't fully understand it. Development moves at a vertiginous pace, faster than anyone outside the major labs can track, and I've given up trying to keep up with every release from every megalomaniacal company competing for attention in Silicon Valley and the handful of places nearby doing the same thing. I mention that upfront because what follows is the view of a daily user still forming his opinion, offered without the authority of someone who builds these systems for a living.

One prediction attached to AI is that it will eventually replace or endanger the human race, usually through some version of machines taking control from the people who built them. I don't find this convincing, at least not yet. Human behaviour is too varied and too unpredictable for a single trajectory like that to hold up as anything more than speculation.

Before going further, it's worth stating plainly how I see the technology right now. AI is a tool. I know that phrase gets repeated everywhere at the moment, practically to the point of self-parody, and I don't mind repeating it again anyway, because it remains the most accurate description available. It will replace the weaker version of human labour that did roughly the same job as the tool itself, but it won't replace professionals who understand the theory behind their discipline, the principles and the science they've spent tens of thousands of hours learning.

This distinction matters because AI can't replace human interaction either. People still need connection, and they still need to delegate responsibility for the tasks around them to other people they trust. Societies are built on exchange and the division of labour, and that division is what produces efficiency in the first place. It isn't disappearing because a new tool entered the picture.

What I've observed in the commercial & marketing space:

The risk I've noticed in digital marketing specifically is how products get made and how people respond to them. Products built entirely from AI output, with no human creative input behind them, are becoming a weakness for brands rather than a shortcut. When a brand's identity disappears because its digital products are built the same way as everyone else's, consumer trust erodes hard. Trust is easy to underestimate. Most brands only notice how much of it they've spent once a customer stops responding to anything they send. No matter how strong your copy is or how well you apply persuasive techniques, a lot of - otherwise profitable - opportunities will slip away once people stop trusting you.

This lack of differentiation produces a specific kind of fatigue. People start to associate a product with every other identical-looking product that has previously let them down, and the association sticks even when the product in front of them is perfectly competent. Pattern recognition matters more here than actual quality. A shopper who has already been burned by a similar-looking site won't stop to judge whether this one is different, because their first instinct is to assume it isn't.

This shows up clearly in ecommerce, which is the market I work in every day. AI can already generate a full storefront, from product copy and imagery to page layout, drawing on whatever the underlying model has seen most often during training. None of that is inherently bad on its own. The problem is that thousands of unrelated stores end up pulling from the same training data and the same default aesthetic choices, so the storefronts start to wear the same facade, brand after brand. A shopper who has been served the same layout and the same headline structure across a dozen unrelated brands stops reading brand as a meaningful signal at all, and that's a harder problem to fix than a badly designed page.

Taking an AI-made product and using it without adding an identity of its own builds a layer of insecurity into the brand, and that insecurity eventually costs money. The answer is to keep using AI while inserting strong elements of identity and a real point of view into whatever it produces.

This lands hardest on designers and brand creators specifically! Designers can't trade their instincts for convenience, and the temptation to do exactly that is only going to grow as the tool improves. They need to keep learning and become well-read enough in their own discipline to impose their own vision on the work, and to use AI only to execute that vision, never to generate it. Imagine designers like Tobias van Schneider or Karri Saarinen building their whole work with AI, from scratch. Or brands like REPRESENT. They would no longer have a community!

None of this requires rejecting the technology or pretending the disruption isn't real. It requires the same thing every previous shift in tools has demanded from people who wanted their work to stay distinctive, which is a point of view strong enough to survive contact with a tool that gives the same output to everyone who asks it the same question.