Unlock Your Creative Potential: Learn to Generate Stunning AI Art For Your Club

On several occasions, I have discussed how useful AI image generators are generally and how great Member Jungle’s JungleVision AI image generator is in particular. They make it easy to create fun, dynamic, engaging visuals for your clubs, emails, newsletters, blogs, pages and events. They are so much better than relying on stock images. 

Due to this, I’m not going to cover that side of things here; I’m going to take it as read that you know why you should be using AI to help your club and why you should be using AI-generated images for your club. If you don’t know, check out this blog category on the Member Jungle Learning Centre for everything you need: Artificial Intelligence To Help Your Club. 

Today, we will cover how to get the best possible images out of JungleVision AI with its brand-spanking-new Imagen 3 engine. We are going to cover the following things:  

• How to write the perfect AI image prompts 
• How style tags work
• AI image generation tips and tricks

How To Write The Perfect Image Prompt For JungleVision AI

The key to creating a great AI image prompt is to provide specific and detailed information for the AI. The basic format for writing a prompt like this is to include the Subject, Context, Background, and Style. 

In other words, you should clearly state Who is doing What, Why they are doing it, and Where it's happening. Finally, Style refers to the type of image you want, such as a photo, a drawing, a cartoon, etc. So, in practice, it might look like “A happy older woman knitting a cute sweater in a lounge, a 4k still from a movie.” 

 

 

 

So, let’s break down each section individually to explain what it does, why it is important, and how you can get it right. Let’s do this while creating an image that could be used for a fictional event I will run for my book club. 

Also, it’s worth noting that while I’ll be doing this one step at a time for this demonstration, you usually add all the elements to your prompt at the same time. 

Perfecting The Subject

Your subject will likely be a person, animal or object, and it will be the focal point of your image. It is basically what you are trying to get an image of. For this reason, you need to provide as much context as possible. Don’t just ask for a car; give the AI context, ask for a classic 80s muscle car in pristine condition. The better your description, the better your end result will be. For example, if I just say “A man”, this is my result. 

 

 

It's fine; it's just a man, but this obviously wouldn't be what we were aiming for our event image, so let's add some more details. Basically, here we want to describe our subject, which is a man. Is he old, fat, gruff, bearded, skinny, angry, formal? We need to find some way to describe how this man looks so he starts looking a bit more distinctive. 

In the book we are going to cover in this event, the main characters are attacked by demon-like creatures, so let's try to create a bit of a supernatural influence by asking for "a man with demon horns." 

 

 

Okay, so it’s all gone a little 50 Shades of Red here, but despite this, I think we are on the right track. We could keep tweaking this and adding details until we are happy with it. For example, I could say, “A nice man with demon horns and a large beard, wearing a suit,” and we’d end up with this. 

 

 

But for now, I will stick with a man with demon horns and move on to my context. 

Perfecting The Context

The context is what the person is doing and how they are doing it. It is the action and intention of the subject in your image. It could be as simple as standing and smiling or as complicated as bungee jumping.

For our example, our demon man will be sitting down reading a book, hence my prompt, “A man with demon horns, sitting reading a book.” 

Which gives us this. 

 

 

Again, the more context, the better. Is he sitting on a lounge or a chair on the ground? Is he reading happily, or is he bored? The more descriptors you use, the better the result will be. 

Perfecting The Background

The background of your image is vitally important. Is your car driving quickly along a beach, a mountain road, or a city? The background you give your image will change a lot about it and how people interpret it. 

Back to our book example: The book in question takes place in a Scottish forest, so I will add that to my prompt. A chainsaw is also prominently featured in one scene, so I will add a chainsaw as a set dressing.  

Our prompt is now “a man with demon horns, reading a book while sitting in a forest, there is a chainsaw on the ground nearby,” 

 

You may have noticed that I left Scottish out of my prompt. I did this because I was worried that the AI would try to make it a distinctly Scottish forest by adding a flag or Ewan McGregor to my background, which I don't need in this particular context and only in this context. 

 

 

Also, your background is the right time to think about adding the time of day to your prompt for some added flavour. 

Perfecting The Style

Finally, we come to the style. This is actually simple. All you need to do is tell the AI what kind of image you want. Are you after something photorealistic, cartoony, an oil painting, film noir, or any other style? Everything up until this point describes what is in your image; this is describing the image itself. 

For example, “A 4K still from a movie” is an excellent prompt for creating photorealistic images. I will also add a list of some good style prompts below, but for now, let’s get back to our book example. 

The book mentions tarot cards, so I asked for the image to be in the style of a tarot card: “A man with demon horns, reading a book while sitting in a forest, there is a chainsaw on the ground nearby, in the style of a tarot card.” 

 

Now, I really like this, and it’s almost perfect. But I have one issue: I never really wanted the man to be red. I was kind of hoping for a regular man just with horns. So, I will try adding “black and white” to my prompt.  

 “A man with demon horns, reading a book while sitting in a forest, there is a chainsaw on the ground nearby, in the style of a tarot card, black and white”.

 

 

This is perfect. It's exactly what I was looking for. With a few quick tweaks in Canva, I have the perfect banner image for my new club event. 

 

 

Not only is this a more intriguing and captivating image for my event post than any stock image, but I know for a fact that it isn't under copyright. So, I'm definitely fine using it to advertise my event, something you can't always be sure of with images from Google searches. 

 

Finally, I can add this to my event and start building the rest of it, knowing that I have an awesome cover photo to get my members' attention. 

 

 

Perfecting The Prompt

As I said at the start, you wouldn’t usually write your prompts in sections like I have. I’ve only done that to show how each part works and what it does. The key to writing a good prompt is giving plenty of details but also trial and error. You will probably need to tweak your prompt a fair few times, adding in and removing various details until you get the result you’re after. 

As useful as this tool is, it can also be a lot of fun, so don’t be afraid to play with it and work out what works and what doesn’t. 

How Style Tags Influence Your Results

As we showed above, style tags are a great way to influence the overall style and tone of your image. They can be entered as part of your prompt in plain English, as I did above with my book club example. However, there are a few built-in style tags in JungleVision located on the right-hand side of the generator. 

These do the exact same thing as entering the style in the Prompt bar at the top. The main difference is that these are tried-and-tested styles, and we know they get great results. 

The available preset styles are: 

• Normal
• Photorealistic 
• Anime 
• Black and White
• Claymation
• Amigurumi - which is a style of knitting

 

So, clicking on one of these style options is a great way to start experimenting with your style without having to come up with ideas yourself. 

To show you how this works, I entered the prompt “an old lady running through a playground” and then generated it in several different styles, including all of the presets and, just for fun, the style of the TV show Family Guy. 

 

 

As you can see, this did an awesome job of turning my prompt into six distinctly different styles of images. 

So, play around with the built-in styles and a few of your own because you will find something that matches your club's voice and tone and will be awesome in your club's communications. 

After all, why have "a car on a beach" when you can have "a car on a beach in the style of a postcard,"? 

 

Both are great images, but if that postcard one was used as an event header image, I’d be clicking on it so fast. 

Tips and tricks

I thought I’d quickly share some tips and tricks to help you get the most out of JungleVision AI. 

 

1. On the right-hand side of the JungleVison generator, there is a box where you can enter negative prompts. These are descriptions of things you don’t want to see in your image, so if you keep getting odd elements in your images that you don’t want, try using a negative prompt or two. 
    
2. Try not to generate celebrities; the AI struggles with this. Instead, try their likeness or a character they play. For example, "Emma Watson eating a bagel on the New York Subway" won't generate, but "Hermione Granger eating a bagel on the New York Subway" does.

 

 

 

3. Giving JungleVision additional background information will help it zoom out. Sometimes, when you ask for an image in JungleVision, it will show it too zoomed in, giving the AI more context to help it zoom out and take in the whole picture. 

 

For example, “A unicorn pinata” gives us this odd zoomed-in image of a pinata in a blank void, while “A unicorn pinata hanging from a large tree in a green park. Surrounding the pinata is a party of kids " gives us a much more zoomed-out and better image.

 

4. Don't be afraid to be weird. JungleVision, for example, “A giant hamburger doing a drift with Darth Vader on top,”

 

 

5. Finally, you will get several failed prompts from JungleVision; this is just part of it. This is because there is a constant battle between people who make AI image generators and people who try to misuse them to create inappropriate images. So, image generators are always being tweaked to help ensure that they aren’t used to create any harmful images. 

This often means that the generation guidelines seem a little too strict, and seemingly harmless prompts are rejected due to crossing these guidelines. If what you are to create is truly harmless, you will just need to keep tweaking your prompt, and you’ll get there before too long. It’s also worth noting that prompts that fail for this reason do not cost you anything. 

I know this can be a bit of a pain, but I think we can all agree that tweaking a prompt occasionally is much better than the alternative. 

What Applications Do AI-Generated Images Have For Your Club?

AI-generated images are an amazing and fun way to make captivating, interesting and copyright-free images to add to your club’s website and communications. They boost member engagement and help add to your club’s atmosphere.

For more on their applications, please check out How AI Can Help Your Club or Association - AI-Generated Art.

For more on JungleVision AI in particular, have a look at How To Create AI-Generated Images Directly From Your Club’s Website.