Learn from my worst AI images and fix these biggest AI mistakes

I spend more time using AI image generators than the average person. I’ve reviewed several programs, ranked and crowned the best, and learned the best practices for AI image prompt engineering to get better results. In the process, I’ve generated hundreds of AI images. Some of the images are beautiful and stunningly detailed, while others are terrible. Otherwise, they can be weird, quirky, scary, or otherwise unusable.

Every bad AI image is unique in its own way, but during my testing and review, I’ve found similar issues across several services. There are certain details that AI image generators tend to screw up, and there are certain tricks you can use to fix them. None of these strategies are perfect — they’re highly dependent on the generator you use — but they’re a good starting point.

Remember that no AI art service can copy or replace human creators. If you use AI image generators, these are the biggest problems I found and how to fix them.

Human faces and expressions

Accurate facial expressions are a constant challenge for AI generators. Weird eyes, teeth, and eyebrows are some of the strongest indicators that an image was generated by AI. In this case, the result was extremely funny to me, but also completely unusable. The girls have Halloween-style vampire teeth, and the guy in the back is having more than just a bad hair day.

Even with cartoonish or non-realistic characters, generators struggle to moderate emotions and expressions. This image — created by our top pick, Dall-E 3 — overexposed the prompt, and the end result was overly dramatic. I consider myself a tidy type, but I can’t imagine anyone getting that upset over what looks like hundreds of dollars worth of cleaning supplies. Even the best programs can fail and produce strange results.

How to fix this: I recommend asking the service to reduce the number of people trying to render it — reducing the chances of errors — and using post-generation editing tools to select specific parts of the image to regenerate or repair. Choosing a milder adjective (“angry” instead of “furious”) can help the service find the right path.

Iconic characters, trademarks and recognizable logos

For many reasons, AI image generators can rarely produce accurate logos, well-known trademarks, or otherwise recognizable marks. This is usually because there are significant legal concerns about a company being able to use its branded materials or intellectual property for AI services. It could also be that the logo or mark you want to include is not in the generator’s training data, causing the generator to not understand your request.

Two recent exceptions to this are new image generation features on the Google Pixel 9 and Pixel 9 XL, thanks to Gemini AI, which our mobile experts used to create images with accurate-looking Mickey Mouse and Pikachu renderings. The other exception is some paying X (formerly Twitter) users who report that they can use the social media site’s Grok chatbot to realistic looking Mickeys and human political figures. These images are still not perfect renderings, but they come pretty close.

How to fix it: You can’t, but on the plus side, you’re much more likely to avoid copyright infringement lawsuits. If you run into this problem, I’d suggest re-examining your design concept to see if you can get the same point across without using a specific company name or logo. Do you really need the TikTok logo, or do you just need a phone with a vertical video playing, for example?

Overlapping, complex elements

Sometimes you can crash a generator if you have too many overlapping or complex elements in an image. The best AI image programs won’t encounter this problem very often, but even the best programs fail at some point. With this Leonardo image, the library of my dreams is destroyed by the rolling ladder disappearing halfway through.

Likewise, this photorealistic image of a kitchen looks fine at first glance, but a second glance or a quick zoom reveals gibberish markings printed on the cookbook, and the book itself appears to have two spines and three sections. Images like this show just how frustrating it can be to create AI images — even when you think you’ve got a winner, small flaws make them unusable.

How to fix it: Try simplifying your prompt or using post-generation editing tools to identify and fix the problem. Sometimes changing the aesthetic can also help; these issues usually crop up when trying to create photorealistic or stock photos. If your service offers a specific area editing tool, try using that to select the questionable area and ask the generator to remove what’s in it.

Over-processing and hallucinations

The sign of a great AI image generator is its ability to produce an image with few errors or hallucinations while giving you the tools to fix those problems when they inevitably crop up, but not every service does that. In this example from Midjourney, I ran several rounds of edits on this image of a soccer team celebrating a victory, and I have no idea what went into it to get to this end result — or any idea what’s going on with the blob that’s presumably a player on the right side of the image. Frankly, I don’t think Midjourney knew either.

How to fix it: When editing AI images, sometimes less is more. Don’t be afraid to scrap your current batch of images and start over. You can often fix big problems preemptively by refining your prompt, so you only have to worry about fixing small issues later.

AI generators still need human work

I have no doubt that the companies behind these AI generators are working to solve these problems and improve their models. For now, imperfect images show us once again that while this technology is improving, it is not perfect. I hope that because so many of these services face similar challenges, they will all look for solutions to these problems.

Finally, I want to remind you that it is a best practice to mention or acknowledge that an image is AI-generated when you share it. As AI images get better and more realistic, it is more important than ever to distinguish AI-generated art from other types of media.

For more information, check out our advice for writing the best AI image assignment and our top picks for AI image generators.