EXPLORING THE AI WORLD 
AI image generation, particularly text-to-image generation, is the current hottest visual design trend. The AI technology that makes it possible to create pictures out of the blue just by using text descriptions has now reached the masses with various online tools and rapidly gaining popularity.
The chance to bring images from your imagination into reality by just typing them out expands the non-designer’s creativity. For business owners and designers, it’s an effort and time saver.
If you want to give it a shot and create beautiful images without a camera and without being an experienced artist, we share a list of the best AI image generators you can try out right now!
An AI image generator is a web tool or app that creates images automatically from text prompts. It’s based on visual artificial intelligence, with algorithms that turn text into pictures by analyzing the words in the text prompt, applying machine learning to correlate them with images and graphic elements, and creating an entirely new image from it. Usually, this kind of tool makes multiple images from one single prompt. These AI generative models are called “text to image” art generators.
These innovative image-creation tools allow for making a unique image out of thin air and unleash limitless creativity: You can make a perfect photo even if you don’t know the first thing about photography, or a beautiful oil painting image, even if you can’t even hold a brush.
They also have various practical applications. For designers, AI generated artwork can help with client work, to visualize an idea, and get constructive feedback easier, for example. But most of all, they’re a great resource to save time, effort, and money when sourcing images for marketing, advertising, and other professional needs.
For business owners, marketers, freelance professionals, and eager DIY-ers, creating images with an AI image generator can mean getting all the visual content they need much faster and at a fraction of the cost, or even free, without hiring anyone to do it.
AI-generated art, especially visuals, is slowly but firmly taking over the creative world. If you don't believe us, check out the numbers in the latest visual AI stats.
In AI text-to-image generation, the text that describes the image you want to create –known as text prompt– is essential.
The AI software will base its creation solely on its understanding of this prompt, so you want to make sure you use clear language and describe what you expect to see as accurately as possible. Some AI models include guidelines for creating an excellent prompt; that’s how important it is.
The more precise and elaborate a text prompt is, the higher the chance that the images generated from it are exactly what you envisioned. Terms to include in a prompt are not limited to main objects and scenes but also colors, specific objects or details, and art style (think of photorealistic or photographs, 3D, digital painting, etc.), among others.
Descriptive words that help add context and meaning are also beneficial. If you include humans in your image, it’s good to detail the desired emotion, overall mood, and more information to set the visual scene right.
On the flip side, you can deliberately make your prompt vague or simple to discover what the software comes up with. Just know that the less information you give, the more the machine has to guess; that doesn’t always come out right.
In the end, how good an AI image is, depends a lot on how good the text prompt is. Soon enough, writing good text prompts will be an art in itself.
A Heads Up on Text Prompts – As AI-generated pictures slowly but steadily enter the commercial use realm, more and more AI image creator tools are adding filters and bans to avoid illegal content made with their software, including what you can and cannot put in your prompts.
Some apps do not allow the use of names of real people (especially celebrities), trademarks, private property, or protected landmarks in prompts. Some are even forbidding users from using other artists distinctive styles (like Van Gogh or Dalí, for example). And almost all of them have rules against the use of terms referring to or easily linked to violence, crime, or illegal activities.
Overall, text prompts must be carefully worded to produce original artwork with artistic and commercial value.
This is our list of the best AI image generation tools currently available online, with access to anyone. Some are free, some have a free and a premium version, and some are premium, paid apps. But all are worth exploring and can produce high quality images while making your life easier.
The leading stock media licensing company Shutterstock now has its own AI image generator. It's called Shutterstock Generate, built on the base of the Dall-E model (see more about Dall-E further down in this list). It's a text-to-image generative tool, where you enter a text prompt and the software delivers a batch of synthetic images based on it.
A few things make this tool stand out, and differentiate from Dall-E itself. First of all, it is trained exclusively with content from Shutterstock's libraries. This content is all sourced from artists around the globe, reviewed for legal validity and with all rights cleared for commercial use. Furthermore, the company has a rewards system in place that compensates contributors whenever their work is used in the training of AI software, and in the synthetization of new images. This results in AI-generated visuals that are potentially much safer, legally speaking, than those produced with other similar tools.
Then, it's a very user-friendly application that seeks to simplify your workflow. For this reason, it doesn't require complex text prompts to work. It can deliver images even from single-word prompts. However, the more detailed your description is, the better the results will be.
Plus, Shutterstock Generate is part of the Shutterstock Creative Flow platform and streamlines with its other tools. So you can create an image using the AI app, then edit it to your liking in Shutterstock Create, use the scheduling feature to program its publishing, and more, all in one place.
Shutterstock Generate is free to access and synthesize images, but the results are watermarked and for preview only. If you wish to download and use an image you generated, you need to pay for it. The system is the same as with their stock photos: one image = one download. Image price depends on whether you acquire an image pack or a subscription, but you can expect to pay between $0.26 and $9.80 per image. It's worth noting here that what you're paying for isn't just the image, but a Shutterstock-backed license to use the image in certain established ways. While this doesn't give you copyright ownership of the AI pictures, it does mean you have a bit more of legal backing for using the images commercially.
All in all, Shutterstock Generate is a promising tool for creatives. You can try it right on the Shutterstock Generate page.
Getimg.ai is a platform that offers two different products, both built on AI technology. One is a text generator, and the other is a fantastic text to image creation tool built on the base of Stability.ai’s Stable Diffusion software –which you’ll learn more about if you keep reading below–.
This AI image generator creates visual content from text descriptions with surprising accuracy and a significant level of detail. Each image you create equals one credit spent.
Besides a text field for you to enter your prompt, it includes a “negative prompt” box, where you can add the terms you want to exclude from your image; this sets them aside since almost no other AI image generators have something like this. The field appears pre-filled with the example “disfigured, cartoon, blurry,” which shows the terms you may want to rule out from your pictures.
Another cool feature lets you use an image as the base to create variants. Furthermore, you can adjust the image creation settings for the number of pictures you want to make from each prompt, the image dimensions, the AI model to use, additional tweaks to apply, and more.
Recently, they've launched a very exciting feature, Dreambooth, that lets you create your own AI generative model, using your own images to train it. All you have to do is indicate which subject the model needs to learn (man, woman, car, styles, etc.), then upload 10 to 20 photos of said subject and follow the page's instructions to train what will then become your custom text-to-image generator! You can then use this model on the Getimg.ai platform or download the model files to use it on your own machine. Remember that the better photos you upload, the better the results will be. At this time, you can create an unlimited number of models, and it costs you 5,000 credits each –this is a promotional launching price and it might be adjusted later on.
And there’s even more: the platform hosts an AI image editor to continue adjusting your picture’s details, and it includes an outpaiting function that can turn a small shot into an immersive image.
The website does not claim copyright over the content you create using it. All visuals you generate come with a CreativeML Open RAIL-M license, meaning you get a lot of control over the picture, including commercial use and distribution rights, but with some restrictions. You can read the complete license agreement here.
Getimg.ai is a sign-up service but has a free forever account with 100 credits per month that you can use to generate photos without spending money. You need to pay for an upgrade if you want more than that. There are various tiers, starting at 3,000 credits a month for $12/mo.
One hundred free AI images a month is a lot, though, and adding the great results this tool produces, it’s almost too good to be true!
Generated.Photos have a very transparent name. It’s a website where you can license AI-created photos of people. You can opt for the download of existing images, create your picture, or even license entire datasets to train your software.
Initially, they only offered headshots (faces, basically), but they have recently launched full-body photos too.
They also have additional features, such as an anonymizer that transforms a face of your choice into a fake one that resembles it but isn’t quite the same.
They have a 3-day free trial for the face generator (though the images you create during the test are watermarked) and a free-forever version with low-res pictures for personal use only. Suppose you want to download high-quality and commercially-usable creations. In that case, prices start at $1.99 per photo on demand (discounts for bundle purchases) or $19.99/mo for a subscription that gives you a fixed number of downloads or creations.
The best part is that the company legally guarantees all images, so they’re safer to use than many other options.
Dall-E by OpenAI is an AI tool to create visual art from text prompts. Recently made available for the public, this technology lets you create any visual you can imagine and produce photo-realistic pictures, paintings, digital illustrations, and more.
Additionally, you can also edit or get variations from existing images. You just describe what you want in a text prompt, and the software creates it.
While there isn’t much information about the legal status of the data used to train this software –some research already produced interesting findings–the website ensures you get full copyright and re-selling rights for any image you create with the tool.
Dall-E 2 (the latest version) currently gives you an initial batch of free credits on your account and a smaller number of monthly credits you can use to generate or edit images. And you can also purchase credit packs if you want more, at $15 for 115 credits.
There is a lot of buzz around Dall-E –partially due to Open AI being an Elon Musk development– and recently, it’s been partnering with high-profile companies such as Shutterstock and Microsoft in AI image generation features that will integrate into their services.
It’s certainly a tool you can have a lot of fun with, coming up with crazy visual concepts or trying out a new style for your art. Plus, it has an easy to use interface.
Note: There is also Dall-E Mini (now rebranded Craiyon), a tool with very similar functionality developed by programmer Boris Dayma. The two tools are not related despite their similarities and initially almost identical names.
Midjourney is a text-to-image AI picture generator that is gaining much traction on the web. It lets users create images on their computer from a written description. While it can produce photo-realistic pictures, this tool seems more indicated to create art in various digital illustration styles.
Many artists are praising the distinct artistic quality of the creations this app is capable of, which seems to set it aside from competitors and alternatives listed in this article.
While Midjourney is open for everyone via sign-up on their discord channel, and you can use it for free, the free creations are under Creative Commons Noncommercial 4.0 Attribution International License, which means you can only use them for personal projects. You must credit and link or tag Midjourney along with the image.
According to its Terms of Use, it is possible to “own” assets created with this tool and potentially use them for commercial purposes, as long as you have a paid subscription from at least the Phase 3 payment plan. Their subscription prices start at $10/mo, but not all enable you to use the assets commercially. There are exceptions; for example, if you're a corporation, you need a corporate membership. In any case, you should contact Midjourney directly to verify whether you can or cannot use an asset commercially before the fact.
An interesting point, the terms of service establish that you give the company full ownership and royalty rights to any text and visual content you create on the platform. Something to keep in mind if you intend to profit from your creations. However, they offer certain “privacy” rights as purchasable add-ons for your subscriptions that prevent your assets from being available to everyone on the platform, at least.
Stable Diffusion is an open-source AI, text-to-image software that allows users to create images from a written description.
Stability.ai developed it in collaboration with other organizations and collectives such as Etheureai, LAION, and DreamStudio. As it’s an open tool, developers worldwide have already released new versions and plug-ins based on the original code –including Getimg.ai, which is listed above.
Currently in beta and open to everyone, you can join through DreamStudio to use Stable Diffusion and create your custom images with AI. All users get an initial batch of 200 credits for free, but once those are used up, the only way to keep using the tool is to purchase credits, at $10 packs for ten credits.
However, one thing to keep in mind is that due to the open-source status of the software, all images produced with Stable Diffusion are under a Creative Commons CC0 1.0 Universal license, which is essentially the public domain. In other words: no copyright for you (or for anyone). These images are free to use however you want, and anyone else wants.
Nightcafe Creator is an AI art generator developed by entrepreneur Angus Russell that uses AI technology to generate personalized art in any given style, parting from text input.
It’s available for everyone upon free sign-up, and it has an easy to use interface where you can enter a text description for a visual, wait a few seconds, and the system will create it for you. You can further “evolve” your image until you get what you had in mind. And their guidelines include helpful creation tips.
This app gives all users five free credits per day that are not accumulable. If you want to create more images out of that limit, you can purchase credit packs from $7.99 for 40 credits or get a subscription to Nightcafe Pro, starting at $9.99/mo for 100 credits.
Interestingly, this app also hosts an online community platform for artists to share their creations, give and receive feedback, and has a reward system where you can earn free credits for things like completing your profile, liking a number of images, receiving a number of likes, and more.
The terms of service disclose that, provided you didn’t use any copyrighted images during the creation, you get full copyright ownership of the visuals you make with the tool. However, they also inform that the copyright of AI art is not recognized or even possible in all regions, so it’s subject to your local law.
Now you know which tools to try first, we want to give you some insight into what to look for in an AI image generator when you intend to use the pictures for professional or commercial use –such as marketing, advertising, online branding, client work, and more.
First, the creative industry is fully embracing AI art, and it’s certainly moving toward commercial usage of AI imagery.
From design software powerhouses to prestigious developers to top stock photo agencies, every big company is now working with AI generative models in one way or the other. Some have partnered with an AI developer to integrate their tech into their platforms, others are developing their native tools, and others are adjusting their services’ legal frames to cover or include AI art.
But they’re all coinciding in one thing: the main concern is copyright and making sure the person using a picture is legally clear to do so.
For this reason, it’s imperious that you check the terms of use from your chosen AI image generator and fully understand what rights you get over the content you create with it, in which ways you can use it, and in which ways you can’t.
Never assume that because you paid for the right to use the tool, the resulting images are legally yours to do whatever, for example. Or that because you used a free tool, that means you can’t use the pictures commercially. Always read the fine print before using AI visuals with commercial intent.
Let’s quickly address the most common questions about AI art generators!
For photorealistic people images, Getimg.ai is one of the best. It produces realistic and natural-looking faces from the first attempt.
For visual art in general, Dall-E has a wide range of results, but Midjourney has a signature artistic look.
These are just some examples.
AI image generators work using neural networks. The software is trained to analyze and pair words and phrases with visual elements to construct a new image after a text description, using “scraps” of millions and even billions of existing images.
The image generated depends entirely on the words added to the prompt, so using descriptive and accurate terms is vital.
Not always. Copyright laws worldwide are still adjusting to this new type of media, and there are a lot of questions about the validity of copyright in AI-generated pictures. Some tools, like Dall-E, assure users they have full copyright of the content they create. That doesn’t mean said copyright is valid or acknowledged everywhere. And some tools, like Stable Diffusion, make it clear that the images made with their app are in the public domain.
And there you have it. We hope you can create any visual you can imagine with any of the listed AI art generators in a few simple steps.
And who knows, you may become a famous AI visual artist someday!
THE AUTHOR
Amos Struck
As an ICT expert with over 20 years of experience, I have worked on a number of projects also involving AI technology. I have been writing about image technology related to design and photography since 2006, and have gained a reputation as an expert in the stock photo industry.
AI Secrets is a platform for tech decision-makers to learn about AI technology. Our team includes experts such as Amos Struck (20+ yrs ICT, Stock Photo, AI), Ivanna Attie (expert in digital comms, design, stock media), and more who share their views on AI.
Get AI news in your inbox & join thousands of engineers and managers using AI to boost sales and grow market share
