We’ve written previously about how to use AI for video editing, how to use AI for copywriting and the best AI tools for busy L&D teams… and now we’re turning the spotlight onto the best AI for generating images!
By now, most of us have played around with ChatGPT and DALL-E to create all sorts of weird and wonderful images, but it’s more than just a gimmick – in fact, if you don’t have a dedicated graphic designer on your L&D team, AI can be an absolute lifesaver for creating images that fit seamlessly into the look and feel of your learning programme and wider brands, or simply for inspiration for those days when your creative muscles just aren’t flexing.
So, let’s take a look at some of the ways in which you can use AI to generate images for your learning, as well as some of the top tools learning professionals are using today to create visually appealing, engaging assets.
Zhuzhing up an LMS doesn’t have to mean learning to code to customise every element of the out-of-the-box design. Thumbnails and icons go a long way to making your LMS feel like yours, and you can dramatically change the look and feel of a learning platform with the on-site imagery you choose.
When you use AI for generating images, the world is your oyster. You can ask the tool to create a whole library of thumbnail images and icons to help spotlight your courses and learning assets, all in consistent dimensions and styles. You can choose the colour palette, art style and more, ensuring the images grab your learners’ attention and help boost learner engagement across your LMS.
If you offer scenario-based training, you’ll know that professional photoshoots can be expensive – especially if you need images for every step of every branching scenario. Likewise, commissioning an illustrator to draw each image can be costly. But that doesn’t need to hold you back anymore!
Using AI for generating images means you can tell it what you want to see in each image, then at the click of a button, you can generate the scenarios you want. If things change down the line, you don’t need to recreate the original photoshoot or use outdated images – you can simply generate a replacement. Easy!
Stock images have their place, but the main issue is that they feel generic. If your learners feel like they’ve seen the images before, they won’t grab their attention, and they probably won’t feel relatable. After all, how eye-catching can the typical ‘businesswoman typing on a laptop’ stock image really be in 2025?
AI also helps you fill in gaps in the typical stock image offering. For instance, women are typically underrepresented in stock images for certain careers, so AI helps you more accurately represent your learners.
In web design, ‘furniture’ refers to all the non-content elements that give your site an identity and make it easier to navigate. We’re talking headers, footers, sidebars, call-to-actions and popups. While you most likely had a web agency to help design your company’s main website, you may not have the same support for your LMS, which can leave it feeling generic or bland.
Once again, this is where AI comes to the rescue. AI is great for generating images to grab the attention of learners and help draw their eye to important elements, whether that’s menus, buttons or featured courses. Addressing these little details can take your LMS from out-of-the-box to knock-their-socks-off, and will help you ensure that the whole platform feels more cohesive and professional to elevate the learner experience.
Internal comms for learning can be tough – but it’s super important you get it right to maintain high levels of learner engagement. This can be a great opportunity to be a bit playful or try different designs to pique the interest of your learners (much like Thomas Thor did when they reeled in learners with a curious image of Ted Lasso’s moustache).
Maybe you could experiment with weaving your SME’s headshot into a relevant scene, or create a funny scenario based on the course content. You could even create a little game – e.g. a Where’s Wally-esque scene where the learner has to spot the health and safety hazards to drum up interest in an otherwise dry compliance course, or a spot-the-difference pair of images to draw learners in.
For each of the following tools, we put them to the test with the same prompt:
“Create an image showing a female train station employee giving a male passenger directions on a train platform at Clapham Junction. The employee should have black hair and be wearing a hi-vis jacket, and the passenger should be wearing glasses and carrying a briefcase.”
Here are the best results we got from the tools we tested...
Current reviews suggest that DALL-E is the best AI for generating images in 2025, owing to its ease of use and the quality of the output. Of course, as with all AI tools, the quality and accuracy of the image generated depends on the user’s prompt, but DALL-E is generally considered one of the most reliable AI image generators.
You can access DALL-E through your normal ChatGPT account, making it an easy tool to play around with if you’re new to the world of AI image generators, and as an added bonus, it will store your AI images alongside your ChatGPT text conversations.
Leonardo.Ai describes itself as ‘the ultimate creative companion’, which is probably why it’s so popular with AI artists. But while you can use Leonardo to design your own video game characters or fantasy worlds, it’s just as effective at generating AI images for your elearning courses within seconds in a huge range of art styles.
You can iterate on the images generated to refine the details or make swaps. For instance, if you’re building training for a supermarket, you can change the uniform colour from green to orange with a simple request, ensuring the artwork resonates with your audience.
Canva is the secret workhorse of many a small L&D team. Now, not only can you use it to design your own images from scratch (or from a template), but you can also generate your own unique images in seconds.
What’s nice about creating your AI images in Canva is that you can keep them together with your other design projects, ready to be dropped in wherever you need them. The new Dream Lab feature allows you to play around with different image styles, so you can generate everything from cartoons to portraits to pop art.
Ideogram just WORKS. It generates multiple AI images at a time, allowing you to pick from a selection, and the quality is fantastic. When you upgrade to a paid plan, you can do things like choose the colour palette or remove the background, giving you even more control over the output, but a quick play with Ideogram will demonstrate how easy it is to generate great images fast.
If you want something more stylised, you can take those images and ‘remix’ them into different styles, including anime or 3D models, helping you switch effortlessly between professional and playful art styles to see which resonates the best with your audience.
Remix works on a credit-based system, allowing you to buy a set number of credits and generate the images you need. As you can see, the free version creates images with watermarks, so you’ll need to pay for clean images, but the quality is solid – no extra fingers or eyes looking in different directions!
Remix is primarily made for smartphones, with Apple and Android apps available, though there is also a web preview version. One useful feature is the ‘Magic Camera’, where you could take photos of colleagues in a range of scenarios, then ask the Magic Camera to put them in uniforms or different environments for a more cost-effective alternative to a professional photoshoot.
AI’s ability to generate images is impressive – but as always with emerging technologies (including the multitude of other AI tools for L&D), it pays to know what to look out for, because it doesn’t always get things 100% right.
Some of the most common issues with AI-generated images include:
If you encounter any inconsistencies in your AI-generated images, there’s a simple fix – just hit that regenerate button to have another go! It’s best to get someone else to review the images to make sure you haven’t missed anything, but you can generate multiple images from the same prompt, so you can just keep going until you get one that works for you.