Like many L&D folks, I’ve been experimenting with ChatGPT image generation. Sometimes, ChatGPT does great! But other times, I find it really frustrating. I’ve been working on a scenario recently for a client project, and I need a series of images of two characters talking together. Midjourney is great for generating images of consistent characters, but only with one character at a time. So, I thought I’d try ChatGPT for this project instead. What I discovered is that ChatGPT’s consistency in characters degrades over multiple generations. I also struggled with the position of the characters; ChatGPT keeps placing characters so they aren’t looking at each other.
As I’ve noted elsewhere, I’m not an expert in AI. My perspective is as a practitioner experimenting with AI to figure out how to improve my own work. Image generation is one of the areas where I’ve spent a lot of time, partly because I often can get better results with AI-generated images for scenarios than with stock libraries. But part of “showing my work” with AI means that I show you when it’s not working too.
I’ll show you my results from experimenting with ChatGPT image generation, including what didn’t work and how I finally got better results.
Generating the characters and setting the scene
For this scenario, I have two characters: a graduate student and a professor. They’re meeting in a conference room or classroom with a whiteboard in the background.
I started by generating a single character, the grad student. So far, so good. This is a usable image.
Generating an image of this same character talking worked OK too. The character and scene consistency is pretty good.
Adding a second character
Adding a second character, the professor, is where things got trickier. I prompted ChatGPT to change the angle to a side view so we can see the other person in the conversation.
I prompted for the other character to be sitting on the “opposite side of the table,” but it set them up around a corner. He should be looking at her while listening, but the body positions are wrong. He’s sort of looking past her, not at her.
I tried to adjust the characters’ positions with this prompt. “Change where they’re sitting so they are on opposite sides of the table, directly facing each other, with more space between them. They should not be at the corner of the table like this. Make the image aspect ratio 16×9.”
As you can see below, it’s still not quite right. They are a little farther apart, but still not directly opposite. They are turned toward each other a little better though. Plus, the character consistency is a little lower. The student’s facial features have changed slightly, and the professor’s hair is darker and less gray. Still, this might be good enough for a scenario.
Changing which character is talking
Then, I tried switching speakers. ChatGPT flipped the position of the speakers rather than keeping them in the same place but changing their poses. Obviously, that won’t work.
I got both characters sitting on the right sides of the table, but neither of these characters is looking at each other. Both characters lose some consistency with each subsequent generation.
When I asked for the characters to turn their bodies so they’re looking directly at each other, both the professor’s hair and the students skin became darker. Plus, he still looks like he’s talking to a third person in the room, rather than the graduate student who is now sitting so close that their arms almost touch.
I tried one more time in this set before giving up. “Turn the professor’s head more so he’s looking at the woman’s face.” As you can see, it didn’t turn his head.
Character inconsistency
Here’s a side-by-side comparison of the graduate student character so the inconsistency is more obvious. I always expect some minor inconsistency in AI image generation; that’s part of the reality of working with these tools. But to have a character shift this much after only a handful of iterations was really disappointing. I think ChatGPT does better with character consistency when you’re only working with a single character. In a scene with even two characters, I don’t think it’s usable if the characters change so much.
I think ChatGPT did somewhat better with consistency in the white male character than it did with the Black woman. The consistency still wasn’t great, but I wonder if part of the issue is that I’m deliberately showing diversity in my characters. We know that AI image generators show bias and tend to exaggerate stereotypes. I think there are general issues with getting character consistency with multiple characters, but I suspect that the underlying bias in the training data may be part of the problem too.
Trying a different approach
Obviously, this approach with image generation wasn’t working. Prompting for images in ChatGPT is different from other tools; you can be more conversational and less precise. This conversational prompting lends itself well to iterating and refining images over multiple attempts (but only if the characters can stay consistent).
I decided to try something different. I generated the initial characters in Midjourney instead of in ChatGPT so I was starting with more detailed images. I think Midjourney is better for more interesting, less generic characters than ChatGPT. I prefer this version of the graduate student character rather than the one I generated in ChatGPT. The professor character still looks a little obviously AI, so I will probably go back and regenerate a base image for that character. It was good enough for this experiment though.
I uploaded both of these reference images to ChatGPT and asked it to put them together in a scene. This is one area where ChatGPT can work well; it can combine and remix existing images into new scenes. There’s some loss of detail in the characters, but I can live with that. However, it still set up the two characters at the corner of a table.
Sketch for prompt
After my previous frustrating experience, I decided to do something different to prompt for the layout I wanted in my image. I did a very quick sketch of the scene and how I wanted the characters positioned.
For my prompt, I used, “That’s not quite right. I want the characters to be viewed from a profile view instead of a 3/4 view. The table should be between them, and they are sitting across from each other on opposite sides of the table. See this attached image for the layout. Use this sketch as a guide, and keep the realistic characters but rotate them to a profile view.”
This approach finally got me the results I wanted. The characters are actually looking at each other!
This also worked for changing which character is speaking. There’s some more inconsistency in characters here, especially the graduate student (note her hair and the black strap for an ID around her neck). But this is definitely more on the right track.
Continuing to experiment
I’ll keep working and experimenting with ChatGPT image generation. I’m still not sure that ChatGPT can keep the characters consistent enough across multiple iterations like I need for this scenario. However, since I started with characters generated in Midjourney, I can also create some images of each character separately in Midjourney.
If Midjourney had an easy way to maintain character consistency with multiple characters, I’d probably just stick with that tool.
Sometimes, ChatGPT is clearly the better image generation tool. Tim Slade showed how he used ChatGPT to generate cutout characters to put on slides. Midjourney can’t do transparent background PNGs yet. I’ve had success with a white background for character images in Midjourney, and that’s fairly easy to edit to remove the background. It does require an extra step though. ChatGPT can do it on its own without the extra editing.
It may be that I need a different workflow for building these scenes with multiple characters. There are some other tools out there like Runway that appear to be able to remix character images into new scenes. Tons of tools are working on character consistency right now since it’s important for storytelling across many fields (marketing, entertainment, etc.). If I can’t get it to work in ChatGPT with the current technology, I think it should be possible in some tool.
If you have had success in generating scenes with multiple characters, let me know what tools you’re using. I’ll add them to my list of tools to test.
Just my two cents, but using real models would maintain character consistency, using a real photographer would make staging a breeze, and using a real room with real props would prevent any of the weird unidentifiable objects AI generates sometimes. And if you don’t have the funds or means to produce the photograph yourself, there are numerous stock websites where people take photos just like this for a living. In my personal opinion, no matter how much a prompt is tweaked, the final result will still look AI-generated. Every AI image has the same qualities, small details that have this perfected look to them, people that don’t quite look like people, unidentifiable shapes, repeated textures, etc. As a graphic designer myself, when I see AI in designs, I make a mental note to stay away from that company. There are still countless companies who do not support or use AI, and those companies will continue to get my business. Otherwise, it’s a slap in the face to those of us who went to school and work day-in and -out making designs / taking photos for clients. I love my job, and having to fix people’s AI flyers they bring to me, is the worst part of my job. I would prefer they came to me first with a blank flyer.
Thanks for reading, just my opinion.
Thanks for commenting, Shae. I hear your frustration.
I’ve been working in learning & development for over 20 years, and I can count on one hand the number of projects where we had the budget to hire models for a custom photo shoot. That just isn’t a realistic solution for corporate training teams.
Stock photo sites typically do not actually have the kinds of photos that we need as instructional designers for building scenarios. I need photos of specific types of people in specific settings doing specific actions. I’ve made do with cutout character images in the past, and then I spend hours assembling them into scenes and backgrounds. I actually prefer vector illustrations for that rather than photos because then at least I can edit them rather than relying on photos in a collection to have what I need. But cutout character images aren’t typically found in major stock libraries; you have to use specialized libraries like eLearning Art for that.
A lot of AI images are very obvious. Can you really always tell though? Test yourself with a quiz like this and see how accurate you are. You may be better than me because of your graphic design background, but I think a lot of people assume that the ChatGPT slop images like the ones in this post and the lousy AI restaurant images or whatever’s on Facebook today are the best the tools are capable of.
I’m curious how you score on this one, for example. Did you get 100% https://realorai.dev/
I think it’s fine to not spend your money at businesses that use sloppy AI for flyers. Personally, I’m not going to penalize a small business for spending money on their employees or other expenses and saving time and effort by making a quick flyer in ChatGPT rather than Canva. Is it obvious that they used ChatGPT with that same marker font against a paint swipe with jagged edges? Of course! But they made a business calculation to save money. A professionally designed flyer would be better quality, no doubt. But for a lot of small businesses, the ROI wouldn’t justify the additional cost required to get that quality.
As for your own work, I also think it’s reasonable to set a rule that customers can’t bring you AI-generated flyers. If they’re coming to a professional, then you should act like a professional and set the boundaries and communicate expectations for how people work with you.
so are you saying then that there is no generative limit, just change your wording to trick the system?
I’m really saying the opposite here. Yes, you can improve your wording for prompts, and you can use other strategies for prompting like using sketches and reference images. Those will improve your results.
However, there’s an upper limit to how good image generation can get with ChatGPT. I wrote this post in May 2025, and ChatGPT’s last image model update is a significant upgrade. While last year I had largely given up on using ChatGPT for any image generation because it was so poor in comparison to other models, the last upgrade actually makes it usable.
I’m still going to use other tools as my first choice rather than ChatGPT. I think Nano Banana is still better at consistency overall, for example.
Part of the challenge is that every image generation tool has limits and weaknesses. One of the key strategies to getting professional results is to use multiple tools, switching as you hit the limits with one tool or another.
Hi, Christy. The sketch is such a great idea! I can’t wait to try it. I was wondering how your prompt for the single image carried over to the two shot. Did you have to repeat the description of the woman? Would you mind sharing the prompts you used? Thanks!
When I restarted in a new chat with reference images from Midjourney, I used this prompt to go from the individual images to the combined scene.
“Create an image: A scene with the characters in these attached images. The characters are in a meeting in a library conference room. The woman is talking and explaining something. The man is listening to her, sitting across from her at a table. Keep the characters consistent, including their clothes, hair, and glasses. Aspect ratio 16:9”
I actually didn’t include that result in the blog post above–it put bookcases in the background because of the word “library.”
This was the next prompt, resulting in the image at the corner of the table with whiteboards behind.
“Can you reposition them so the characters are facing each other across the table? I want the scene set so he is looking at her face. Also, change the background. Instead of books behind them, it’s whiteboards on the walls.”
I didn’t repeat the descriptions of either character; I just used the reference image for the details. It may be helpful to note details that are important or that regularly get missed. In Midjourney, I continue with the description each time, just changing the piece I want changed.
Hi, Christy! Did you upload your sketch into ChatGPT and prompt it – “Can you create a scene of two characters talking using the sketch below?” ?
Basically, yes. The prompt was, “That’s not quite right. I want the characters to be viewed from a profile view instead of a 3/4 view. The table should be between them, and they are sitting across from each other on opposite sides of the table. See this attached image for the layout. Use this sketch as a guide, and keep the realistic characters but rotate them to a profile view.”
I just updated the post to include that prompt so it’s clearer how I used the sketch. I think giving it a reference for the layout is an effective strategy, especially when it seems stuck in a particular layout.