Your video’s thumbnail can make all the difference between going viral and disappearing into obscurity.
The perfect thumbnail will help your clip stand out in recommendations and search results. That’ll draw in viewers – and potential subscribers.
The numbers back this up. According to YouTube, 90% of its most successful videos have custom thumbnails.
But generating a great thumbnail can be tricky. After all, not all of us are professional designers – or want to hire a pro just yet.
Fortunately, AI is here to help. By harnessing some of the amazing AI tools now available online, and mostly free, you can make some decent headway towards a captivating thumbnail.
Here’s how.
Contents
Use a Ready-Made AI Thumbnail Designer
The first option to harness AI to make a YouTube thumbnail is to turn to a ready-made AI thumbnail designer.
AI thumbnail designers do pretty much exactly what it says on the tin. They provide you with a suggestion for a YouTube thumbnail based on a prompt – usually your video’s title and subtitle.
Unlike the more general AI tools we’ll discuss in the next section, AI YouTube thumbnail designers specialize in generating outputs specifically for this platform. It’ll even come in the right dimensions – a convenience that isn’t a given, as you’ll see.
However, most of these tools are fairly basic and only provide you with an initial suggestion. You’ll then have to refine it manually.
Let’s take a look at an example.
One free AI thumbnail designer is thumbnail.ai. Here’s what it looks like:
As a prompt, you enter your video’s title and, optionally, its subtitle.
In this case, the initial results are underwhelming. You can improve them by choosing different themes, changing fonts, and modifying the background images. Still, it’s a very basic tool.
Let’s take a look at a more sophisticated process.
Use Specialized AI Tools to Design a YouTube Thumbnail
The alternative to using a ready-made AI thumbnail designer is to combine different AI tools that are available online. We suggest using ChatGPT and Midjourney or DALL·E 2.
Let’s quickly introduce these three.
You’ve probably heard of ChatGPT. It’s probably the best-known generative AI tools out there right now. It was designed by OpenAI to generate human-like text responses in natural language processing tasks such as question answering, translation, and text completion.
Midjourney is an AI image generation tool that produces unique images based on text prompts. Currently, it’s only accessible via the Discord bot, and it’s free to try for the first 25 images. After that, you’ll have to pay $10/month, though.
Finally, there’s DALL·E 2, another recent AI system that can create realistic images based on natural language prompts. Like ChatGPT, it’s made by OpenAI and is free to use once you’ve created an account.
So how can you combine these tools to make an amazing YouTube thumbnail? Here’s the full process we came up with.
Generate Thumbnail Design Ideas With ChatGPT
To start with, use ChatGPT to generate suggestions for thumbnail designs based on your video’s title or content description.
Let’s say you’ve made a clip about using AI to make a YouTube thumbnail. Go to ChatGPT and tell it to: “Give me 5 different thumbnail design ideas for a YouTube video titled ‘Using AI to Make a YouTube Thumbnail’ ”.
You’ll get something like these descriptions in return:
1. A split-screen thumbnail featuring an image of a person holding a camera on one side and an AI interface on the other side, with the title in bold letters in the middle.
2. A thumbnail with a futuristic design showing an AI robot creating a thumbnail, with the YouTube logo in the background and the title in large font at the bottom.
3. A thumbnail with a side-by-side comparison of two different images, one created by AI and one created by a human, with the title in a speech bubble above the images.
4. A thumbnail with a zoomed-in shot of an AI-generated thumbnail, with a magnifying glass and the title in bold letters overlaid on top of the image.
5. A thumbnail with a before-and-after comparison of a thumbnail made without AI and one made with AI, with the title in bold letters and a line separating the two images.
Here’s how that looks like on ChatGPT:
That gives us – or rather the next AI tool – some raw material to work with.
Have Midjourney Draw Up Designs Based on ChatGPT’s Prompts
As a next step, go to Midjourney and have it create an initial suggestion based on one of the prompts from ChatGPT.
For example, let’s refine the second prompt and tell Midjourney to generate: “A futuristic design showing an AI robot creating a YouTube thumbnail, with the YouTube logo in the background –ar 16:9.” The “-ar 16:9” part of the prompt sets the aspect ratio to match YouTube’s thumbnail format.
Let’s take a look at the outputs:
Not bad! But can another tool maybe do better?
Generate Alternative Suggestions with DALL·E 2
To compare, feed the same prompt to DALL–E. This gives you some solid alternatives and a feeling for the capabilities of these two tools.
One major downside that DALL-E currently has is that it only generates square images. While it is amazing at refining images and altering and improving parts of them, this still means that you’ll have to change the aspect ratio of the thumbnail before you can upload it to YouTube.
There is a feature to extend the generated frames in any direction by adding extra squares, called outpainting – but it’s rather fiddly, meaning you likely won’t get the dimensions perfectly right.
Still – you’ll probably have to spruce up the output anyways.
Polish in Photoshop, GIMP or Canva
Finally, you’ll need to refine the images you’ve generated with Midjourney and DALL-E 2. For the moment, these amazing AIs have a few restrictions, such as the limited aspect ratio in DALL-E 2. Image AIs are also notorious for producing distorted faces and hands.
In addition, both of them are pretty limited when it comes to handling text. Given that, it’s likely a good thing that ChatGPT didn’t even try to reproduce the video title we gave it.
To finalize your thumbnails, you can use software like Photoshop or, if you’re on a budget, GIMP or Canva. GIMP is a software that offers very similar features as Photoshop. It’s free and compatible with both Windows and Mac.
Canva is an online graphic design tool that already comes with a wide range of fonts, graphics, and royalty-free images. In the example below, we used Canva to adjust the dimensions of the images and add text to match the video title.
And that’s a wrap! While you can still add more effects and visual pizzazz, these are two solid YouTube thumbnails generated with the help of AI.
Final Thoughts on AI Thumbnailing
For the moment, we’re not quite at the point where an AI can independently draw up the perfect YouTube thumbnail. But it can be an amazing tool to help you generate the perfect thumbnail design.
ChatGPT can quickly suggest a variety of prompts for creative YouTube thumbnails. These prompts can serve as input for graphic AIs like DALL-E and Midjourney. Once you’ve refined the output from these platforms, you’ll have clickable, unique YouTube thumbnails to draw in viewers and supercharge your channel growth.
You can use AIs like Midjourney and DALL-E 2 to generate suggestions for YouTube thumbnails. However, you’ll have to adjust the image dimensions and add text manually using tools like Photoshop, GIMP, or Canva.
Start by having ChatGPT generate prompts for thumbnail designs based on your video title. Then, feed these prompts to graphic AIs like Midjourney and DALL-E 2. Finally, adjust the dimensions of the output and add text using tools like Photoshop, GIMP, or Canva.
Yes. You can give Midjourney a text prompt for a YouTube thumbnail and define the right aspect ratio (“–ar 16:9”). However, you’ll have to edit the output manually if you want to add text.
Yes. You can use AI generated by platforms like Midjourney and DALL-E 2 to serve as the basis of your YouTube thumbnails.
