How to Hack AI Image Generators

What do you think is the most popular use case for AI after Large Language Models? It’s image generators! So much time and money is saved by using an image generator instead of hiring a designer or photographer. And where there have been limits before, our fantasies can go berserk alongside Midjourney, Leonardo.AI, and others and create stunning AI art.

To be fair, more often than not, you will have created a beautiful picture, but just not really what you wanted. So, how do you get the most out of image generators and get them to do what YOU envision? And what are the best tools for the various image generator use cases?

In this blog, I shall share my ultimate AI art prompt tips with you. On my homepage, www.aitoolssme.com, I review AI tools and have compared 15 different AI image generators, giving them the same prompts. This gave me a really good overview and made comparing the different AI models so much easier. And since I've been following this space closely for over one year now, I feel like the perfect person to answer the questions I've raised above.

Common Use Cases for Image Generators and The Best Tool for It

Social Media Images

Since a social media post with pictures will perform much better than a post only with text, it totally makes sense to quickly generate an image that matches the topic. Tools like Leonardo.AI might not be the most premium tools, but their offer comes with generous 25 free tokens. If you are posting on social media daily, Leonardo.AI will have you covered for almost one month for free.

Website

Similar to social media, every blog post and web page with an eye-catching image will draw the audience more in. Due to a certain permanence, images for websites require a more professional approach than social media. Many header banners will consist of a mix of a headline and an image. I recommend using Canva for it, because this popular design tool lets you add layers with text over the generated image in the same workflow.

Image-to-Video / Ad

Let's say you want to run video ads and have to bring a certain image or image style to life via a moving picture. Since you'll have to invest money in the media buying, it doesn't make sense to be cheap on the pictures and end up with a low-quality video. If I wanted to make sure that my images are of premium quality, I'd go with Adobe Firefly. Adobe has understood that to stay on top of the AI game, they have to offer more than just the standard styles every tool offers. The most relevant feature from Adobe Firefly for creating image-to-video is probably the camera angle setting, which lets you change perspectives, which is optimal for a video ad.

Creative Art

If you like to think of yourself as a young Salvador Dali or Frida Kahlo, just for digital art, I have the right tool for you! The old classic, Midjourney, was the first image generator I've ever heard of. It has erased its barrier to entry, the discord login, and is known to generate the best results of visualised fantastic worlds. Plus, I like to call it the ‘creatives' muse’ because every creative director has mentioned Midjourney in their interview for my series 'Spilling the Digital Beans on AI.' Without an exception.

And last but not least Midjourney has a feature called 'character consistency' that helps you generate various images with the exact same look of the protagonist. This was an early problem with image generators: you could just never replicate a style or character. Now you can also re-create a style by adding the sref (style reference number).

Presentations

Trying to convince your boss to give you a higher budget next quarter? What could help is a very convincing presentation with very convincing images supporting your cause. Since you don't have to show your creative skills in this case, but simple images will do, I recommend a solid tool such as Freepik. The 20 free credits should allow you to add some nice complimentary images and drive that higher budget for the next quarter home. These are the most popular use cases for AI to generate images. But before you sit down to prompt away, I'll have one more pigeon up my magical sleeve:

Formula For Writing Effective AI Art Prompts

The basic formula for every image prompt is:

Style + Subject

For more specific requests, I recommend adding

+ Background + Color Palette + Additional Elements

An example : Style: "realistic" , Subject: "a golden retriever" , Background: "in a sunny park" , Color Palette: "warm and vibrant" , Additional Elements: "with soft shadows and a bright blue sky"

Example Prompt: "Create a realistic image of a golden retriever in a sunny park, using a warm and vibrant color palette, with soft shadows and a bright blue sky."

Now I can only leave you with a fair warning: AI Image generators are like time machines! Just like social media or YouTube, you might intend to stick around only for a while. But the first time you raise your head from the screen, you realize how it has become dark outside, and your computer is 134 generated images heavier. Don't say I didn’t warn you!

Lili, a video marketing expert based in Berlin, calls ChatGPT her best friend. On aitoolssme.com she provides non-techie, user-friendly reviews for AI tools especially helpful for small businesses.