AI images are quick and easy to make. They look great at first sight, and they quickly replaced the use of stock images. But, like everything cheap and easy, they come with trade-offs. A short critique of the pure AI image.

Using AI well can lead to great visuals. But, quick AI shots often miss realness, depth, and originality. They can hurt your content’s quality and trust. Here are five times two AI clichés, why we should avoid them, and what we should do instead.

1. We recognize them

When Jurassic Park first hit the cinema, we had a hard time spotting CGI. These days, Jurassic Park looks like a practical joke. People can usually tell when an image is AI-generated even with the most advanced CGI. There’s always an air of PS5 game cinematics.

The Aphex Twin Fan Art: Somewhat funny, but the warped anatomy and the inhumane humor of the synthetic repetitive collage make it feel like an Aphex Twin fan art cover.

The Original: Aphex Twin figured out the AI aesthetic before AI. This is just one of many examples. The photo shows the backside of his famous Window Licker Album (1997). It’s noteworthy that some of his composing techniques share similarities with AI processing.

In 2024, freaky collages don’t make you look like a music artist with visionary ideas, way ahead of his time. This type of imagery is just ordinary cheap AI crap. It got old real quick.

Cheese face that looks like candy: AI becomes recognizable when it mixes different categories that do not belong together. Done intentionally, with an idea it can be refreshing, done randomly it’s just robot stuff.

The robot sugar high: Some images look like they were made under a robotic sugar high. Lots of warm colors, but they make everything look like candy… they’re so overly sweet that they give you visual diabetes..

Average AI images drag down everything around them. An AI hero image is a comedian opening the show with a knock-knock joke. Good images enrich your article, bad images steal its soul.

AI images make your audience think: “If they use cheap AI for images, they probably use it for the rest, too.” It raises questions about the authenticity of your content and your character in general.

“AI Art” escalated quickly. It’s January. While some executives in Davos may get excited about its infinite possibilities this week, to a younger consumer AI Art is already ‘a bit cringe’.

2. They all look the same

If you have the slightest sense of visual language at all, most AI images now look the same to you. They may still seem aesthetically appealing, somewhat, technically, at first, but they are still quickly recognizable as AI images. And then that’s all they communicate now: “I.M.A.I.”

The Pulp Cover: This screams AI. 50’s pulp fiction style with the sexy lady and the creepy dude leeching. AI or not, in 2024 an outright terrible choice.

The original: Wild is the woman. It looks like, AI got a lot of training on 50s and 60s pulp covers. Obviously, the AI picture has a higher resolution, stronger colors, and eventually, it mixes in some more contemporary Instagram model makeup.

The Laser Rainbow: Easily recognizable as AI because of the laser colors and the messed up letters. Yeah, one year ago, this might have looked cool, but now… just don’t.

Laser Clone: There are thousands of those. If something is cheap to do it has low economic and, in an attention economy, information value—even if the same image, previously would have cost some time and money to make it.

Using AI images makes you drown in a sea of sameness. So, to stand out, rather use a grainy self-shot image? The question is: Do you always need an image? The use of images shouldn’t be an unconditional “always.” Instead, it should be a deliberate choice, driven by the specific needs and objectives of your content.

3. They lack emotion

AI images can miss the emotional or detailed touch that real photos or custom graphics have. Often they just feel outright creepy.

The Body Snatcher: Again a typical case. It could be seen as funny but if you’re honest it just isn’t. Despite all the colors, it feels insane in a Body Snatcher kind of way. People do something that they would only do under the influence of an alien invader, that took over their bodies.

The Original: Body Snatchers is a since fiction that has been adapted multiple times. The story: Aliens make clones of real people, but the clones have no feelings. Bit by bit all people are replaced by soulless lookalikes. A bit like AI.

The Carcinogenic: Computers don’t feel, and that means: they don’t understand what they do, they grow images like cancer grows cells: They just replicate something into the blue. This becomes apparent in the often outright creepiness of AI images. A lot of AI images we wouldn’t even think of—unless we wanted to frighten others.

Random Horror: AI is really good at making scary images. Even if the prompt lacks all hints of horror kitsch, you need to get ready to see or feel something disturbing when you look at AI images. It’s like a spell. Part of the scariness comes from the cancer-like pattern that reproduces the same ornament without considering its meaning and consequence.

Images are not just a splash of colors. They tell a story through the objects and their arrangement, and they tell a story through the style and tone. Every detail matters. If the creepiness oozes out of the image, it will affect everything in your communication.

4. Ethical issues

Using AI images can raise questions about cultural bias, authenticity, and copyright. You may or you may not care about those. They are nevertheless very real legal and ethical matters worth thinking about.

The Copyright Issue: A lot of images use stereotypical figures. Beyond the usual eeriness of most AI images, the recognizable coloring, and the anatomic oddities, this picture may make you smile. While Disney may see a problem with you using Indiana Jones, this should not be your primary concern. If you care about individual artists, you might give the careless use of AI imagery a second thought.

The White Kitsch: You need to be aware that there are more pictures of white people on the Internet, and thus there will be more white people in AI images unless you tell it specifically to pick what you envision. Even then you need to pay attention to what the image says. Computers don’t feel what they do.

AI images might look technically appealing at first sight, but they fall apart on closer inspection: AI doesn’t know what it does. If you don’t curate your prompts like Shakespeare wrote his sonnets, you’ll end up with images that don’t say anything but: “I used AI to add color.”

The van Knock-Off: One of the lamest ways to use AI is to generate images in the style of painting artists. This was done early on, and while we may have nodded at first, believe it or not, we’re completely over it now.

The Black-and-White: One way to hide that it’s an AI picture is to prompt it to make black-and-white images. Now, the color scheme won’t give it away anymore, but we can still tell.

5. What about the future?

At this point, our average AI bro, reborn from Cryptohell, will chime in and teach us that this is just the beginning and that very soon, we won’t be able to tell the difference anymore. Well played, AI bro! But you’re off-topic. We can tell now and we need to not fool ourselves that others won’t notice.

The Can-You-Tell: Not all AI images are directly recognizable. This one is well done. You can still see it in the texture’s details, the strap of the guy’s backpack, and the continuity of the lanyard. With a bit of Photoshop, you could hide those issues, but it might still feel like a computer game on a PS6 somehow.

Take a step back: While there may be some great exceptions, we have quickly developed an ability to sense AI-generated images.

Even if AI improves and we can’t tell the difference, we still need to put effort into choosing images that work with our message. This shows that AI, like stock images, can’t replace thoughtful selection in communication.

Now what?

While an obvious stock image says “I’m a cliché”, there is not a lot of Art in most ‘AI art’ either. A directly recognizable AI image just says “R.O.B.O.T.” Obviously, there are hypnotizing and sad stock images, amazing and horrifying AI renders. AI images are generally easy and quick to make but have significant downsides. However, it’s not about whether stock images are generally better than AI. It’s about how quickly AI images have become old and made stock images look even older.

The advent of AI should not be seen as a challenge. AI Art has shown that if we want to be heard and seen we need to show that we care for humans as humans. This has happened before. It was the invention of photography that made us come up with impressionism and, in consequence, modern art. Photography has made us question traditional art. Similarly, AI can make us question empty off-the-shelf communication. Ironically, machine-generated content might catalyze a fresh wave of humane creativity and hand-crafted innovation in verbal and visual storytelling.

Conclusion

AI or not, images should be given the same attention as text. Choose each visual carefully and place it in the right context. Don’t let AI do your job, use it as a tool. Ask yourself what your image means and how it adds to your message. This balance ensures both your words and visuals are powerful and meaningful. Images tell stories. Bad images tell lame stories. A good image tells a good story.