AI IMAGE GENERATORS, TECHNOLOGY, AND DIGITAL MARKETING
How to Get the Same Face Twice with AI Image Generators
7 Tips for rendering visually consistent characters in AI images
For tons of advice from myself and others, visit the Jasper Facebook group!
Creating a strong brand identity is key to success for any business. Your customers should recognize your brand’s “face” across all channels and touchpoints, whether it’s on your website, print media, social media, or a chatbot. And one of the most important aspects of having a brand identity across all of your platforms is using the same person in different images.




This can be a challenge if you’re using AI image generators commercially. The images produced by AI tools can be vastly different from one another, even if they’re using the same input data. So how do you make sure that your AI-generated images remain consistent across your brand identity?
The Number One request my marketing clients have about AI art is:
“How can Artificial Intelligence generate the same person for me again?”
My answer has some professional “special sauce” that I’d like to keep proprietary for those customers. But here are some tips I can share:
1. Use the same input data for your AI image generator
If you use different input data, you’ll get different results. So if you want to generate consistent results, use the same text prompts, photo, or set of photos each time. Save your favorite prompts in a spreadsheet to reuse!


2. Learn to use Seeds correctly
In Midjourney, you can add this parameter to your prompt:
::sameseed
This will make each of the results in your grid use the same “slice” of the seed, and it helps ensure that the output images are similar to each other.
Seeds, btw, are the random noise that is used to initialize the AI’s creation of the image. Here’s my rough-and-ready explanation of how it all works:
You can also set the seed to the number used in a previous result that you would like to replicate (between 0 and 4294967295). Otherwise, it’s random.
3. Use the same medium and style for your AI images
If you use different mediums, you’ll get different results. Here’s my gallery of mediums. So if you want to create a consistent model for your brand, use the same style and medium each time. For example, if you’re wanting a line drawing, you will need to repeat the same medium and artistic style again.


4. Use constraints, creatively!
Limit the possible outputs from your AI image generator with description. For example, specify a distinctive facial feature or an iconic hairstyle. By doing this, you’ll reduce the chances of getting radically different results.
5. Use celebrities
Okay, while many AI generators frown on (or flatly ban) celebrity photos as input or names as text prompts, there are a couple — like Stable Diffusion — that will let you. If you’re looking for consistency, this is a way to achieve it.
Another option is to alter the celebrity photo first so that it’s not an exact likeness of the person, but a stylized version (you could use a painting filter, for example). Or describe the celebrity you want in the image to the AI generator in detail. Some celebrities are easier to describe than others:




