When AI Starts Making Everything Look the Same

I guess I’m not alone in noticing the rise of AI-generated posters, both online and in the real world, advertising local events and businesses.

When I first started seeing them, they certainly caught my eye. They looked impressive. The images were polished, colourful and often far more elaborate than anything a small organisation might ordinarily have been able to produce.

And I get it. If you are a small organisation, AI offers a quick and inexpensive way of creating something that you otherwise might never have been able to produce. It gives you access to a level of visual production that previously required some combination of money, equipment and professional skills.

But now they are everywhere. To the point where, for me anyway, they have lost much of their originality and do little to encourage me to attend the event or visit the business being advertised. Instead, they are beginning to look like a visual style in their own right, and one that is already becoming rather tired.

It also raises some real moral and potentially legal concerns about their use.

My concern is what this approach could lead to. I know many creatives argue that the proliferation of AI-generated content is harming our industry, and I share those concerns. Work that was previously done by a human and supported the creative economy is undoubtedly being replaced by content that can be produced in seconds.

But, in fairness, the content being produced by these hyper-local, small businesses and events was probably never part of the professional creative industry in the first place. I’m talking about small independents who simply made their own advertising content in the best way they could. These are people who can't afford to hire a designer or a photographer for their advertising, and the event or product they are promoting probably wouldn't generate enough income to justify a proper advertising budget.

It is work that creative professionals never got to see in the first place.

So I don't honestly think the AI content I’m talking about is diverting work away from designers and photographers, at least not directly. My concern is what happens if this becomes normalised.

The more AI-generated imagery we see, the more acceptable it becomes. Eventually larger organisations, which do have advertising budgets and employ designers, photographers and marketing people, may start asking why they should spend money on professional creative work when AI can produce something that appears to do the same job, more quickly and at a fraction of the cost.

That, to me, is the real threat to the creative industry. Not necessarily what small independents are doing now, but where their use of AI could lead. It could also result in a gradual watering down of quality and originality, as organisations become accustomed to accepting something that looks good enough rather than something that has actually been thought about.

Professional designers and photographers are adept at communicating a message visually. They understand that creating an effective piece of work isn't simply about making a pretty picture. It requires an understanding of visual language, what the message needs to say, who the audience is and how the image needs to work in order to communicate with that audience. The visual choices are deliberate.

This is what is usually missing from quickly generated AI content. It is worth saying that AI itself isn't necessarily the problem here. AI tends to work best when it is given a well-crafted, considered prompt. You get out what you put in.

If the person creating the advert hasn't thought carefully about the place, the audience, the message or what makes their event or business distinctive, then the AI has very little to work with beyond the broad subject. The result is inevitably bland, with an unsophisticated use of visual language. The subject may be deeply rooted in a particular location or person, but what is produced simply ticks the basic boxes of visual appeal while lacking any real relevance to, or personality of, the product, place or person.

The irony is that something that is intensely local ends up with an image that could have been made for almost anywhere or anyone.

As someone once told me: “Your job isn’t at threat from AI, but your job could be at threat from someone who knows how to use AI.”

There is also a growing reaction against the aesthetic itself. People are becoming increasingly good at recognising obvious AI-generated imagery, and some are now actively put off by it, myself included. What might once have looked impressive can instead look cheap, generic and impersonal. It reminds me of ClipArt in the 1990s: once you had seen enough of it, you could spot it immediately.

Then there is the use of photorealistic AI imagery, which I find equally troubling.

When we see an image that presents itself as a photograph, we naturally assume that what we are looking at is a photograph and that the subject existed in the real world. The people are real, the venue is as it appears, the event happened and the product looks like that.

If that image is then used in advertising without making its artificial nature clear, it raises a different set of concerns. Under the UK's current consumer protection rules, advertising must not create a misleading impression, and the ASA specifically considers the imagery used in an advert when assessing the overall impression it gives consumers. The law also covers an overall presentation that is likely to deceive the average consumer where that is likely to affect their decision.

That doesn't mean that using an AI-generated image is automatically illegal. The problem is what the image communicates. If a photorealistic image gives the impression that a venue, product, person or experience is something that it isn't, then the fact that the image was generated by AI doesn't necessarily protect the advertiser from the consequences of creating that misleading impression.

There is also the reputational damage this could lead to. Realising that what you saw online isn't what you get leaves you with a feeling that you've been deceived. And, regardless of whether there was an intention to deceive, that is how the customer is likely to feel.

And of course there’s the well documented problem AI has with its gender, age, race and disability bias when generating images of people. Successful people are represented as white men in their 40’s, women are sexualised and people with disabilities are almost non-existent in the AI generated world.

I suspect that in time more and more people will tire of the AI-generated illustrative advertising and it will become dated. We are already seeing that happen. What initially looked impressive increasingly looks generic.

But I hope that when it comes to photorealistic AI images, the law and future case law will put clearer boundaries around their use, particularly where they are presented in a way that leads people to believe they are looking at something real.

Because ultimately, advertising is supposed to tell us something about what we are being offered.

If the image is invented, the question is whether the invention is simply creative interpretation, or whether it is creating an impression that isn't true.

And I think that distinction is going to become increasingly important.