If you were alive in the early 2000s, you might remember Hollywood's obsession with artificial intelligence. Today's AI applications are a far cry from the vengeful robots of AI, Robot or the awkward AI-generated child from Spielberg's AI. Now creatives have found the next big thing in AI: the AI art generator.
Unless you've been living under a rock for the past few weeks, you've probably seen AI-generated portraits filling your social media feeds. From the portraits of the historical eras from MyHeritage to the stunning, dreamy headshots from Lensa, it's hard to escape the striking visuals taking the internet by storm.
Can you believe we live in a world where AI makes art from text?
At Worklife, we've had our eye on AI for a while and started leading pre-seed rounds in the space back in 2020. We believe artificial intelligence has an incredible opportunity to enable and empower the creator economy and take art to new levels of the imagination. And while we're still in the beginning stages of Generative AI, we're excited to see what this new generation of creators using AI tools can do.
The future of AI is bright, but it will also bring new challenges. So it's important to consider how AI will change the role of the creator in society, and how we can address these issues to ensure that AI benefits us all.
What is AI-Generated Art?
So what exactly is AI-generated art? And what can it do for the creator economy?
Stability AI founder, Emad Mostaque, says it's the key to unlocking creativity.
"So much of the world is creatively constipated, and we're going to make it so that they can poop rainbows," he said in a New York Times article.
AI art is any artwork created through the use of artificial intelligence algorithms. It's one of computer science's early applications in the creative realm.
To break it down further, artificial intelligence is comprised of complex algorithms designed to mimic the human mind. Essentially, if your mind can make it up, an AI art generator can produce a visual for it.
And while we're seeing a skyrocketing adoption rate by the masses in smartphone apps like LensAI and Dream, there remains a lot of debate surrounding AI-generated art in the creative community itself.
Some of the clear benefits of AI-generated art include the following:
- Continued innovation and evolution
- Generation of realistic or hyper-realistic data
- Impossible becomes possible
- A step towards inclusive design
With the power of AI, creators can generate high-quality visuals at a higher quantity rate, leaving more time for creative thinking–instead of creative doing. As a result, projects can move faster, and visuals can be produced easier and accurately. The list goes on.
Throughout this post, we'll discuss how AI can unleash a whole new world of creativity and innovation in any industry.
On the other hand, there are some clear challenges to AI-generated art, including:
- A lack of human touch
- Potential for repetitive output
- Lack of control over the output
- Ethical concerns
As with any technological advancement, there's a growing fear that artificial intelligence can make jobs obsolete. We saw this same concern with the introduction of robotics in manufacturing. Yet, we've only seen manufacturing jobs transform into new skill sets. And as we saw with the job market during the pandemic, those jobs are still vital to manufacturing facility success.
In this article, we'll also touch on how these challenges are being considered and overcome by leaders in tech.
The Evolution of AI Art and AI Art Generators
As with any technology, AI art had a slow growth before its meteoric rise to fame.
Technically, AI-generated art has been around as long as the early days of computer graphics. In the 1950s and 1960s, algorithms were used to create patterns rendered on a computer screen. Then, in the 1970s and 1980s, we saw the rising popularity of computer-aided design (CAD) programs.
While CAD software has developed over time, going from 2-dimensional floor plans to 3D models made for augmented reality, its adoption was seamless compared to today's AI-generated art tools.
In 2015, the real art history of artificial intelligence began when Google released DeepDream. Remember those surreal, sometimes creepy, images not even a human mind could imagine? At times it felt like the world stepped into a bad acid trip, but it was hard to ignore the promise of deep machine learning and neural networks.
AI-generated art reached a new milestone in 2018 when Portrait de Edmond de Belamy, an AI-generated piece of artwork by French collective Obvious, was sold at an auction for over $432,000. The final price was 43x its estimated worth!
To quote a 2019 article from The Atlantic, "the ai-art gold rush is here."
Today, galleries and art museums across the country showcase AI-generated art. Activists are using AI art to fight for justice. Investors are flocking to get a piece of AI art tech companies like Midjourney, Stable Diffusion, and DALL-E. Creatives are jumpstarting their projects with text generators and generating graphics in minutes.
AI art is here. And it's not going anywhere anytime soon.
How AI Art Generators Work
As humans, we need help understanding what makes something creative. But AI can learn from its environments and create new things we've never seen before. This has led many people to believe that computers can be creative—and maybe even more creative than humans.
AI creates art that looks like it was produced by humans, but the process behind each piece is entirely automated. It's designed to mimic human behavior to create original content without having to learn from scratch. This allows AI to produce art at an unprecedented rate by using existing resources as inspiration.
In the world of AI art generators, there are two main models:
- Neural Style Transfer (NST)
- Generative Adversarial Network (GAN).
Neural Style Transfer (NST) AI Art Generators
Neural Style Transfer (NST) art is not to be confused with NFTs (non-fungible tokens). This type of AI refers to software algorithms that manipulate existing digital images or videos to adopt another image's visual style. Essentially, you combine one base image with another image or set of images to transform the image using deep neural networks and machine learning.
If you've ever used mobile apps like DeepArt and Prisma, you're using an NST AI art generator.
There are limitations to NST art in that you still need a base image and reference images to work with.
Generative Adversarial Network (GAN) AI Art Generators
On the flip side, Generative Adversarial Network (GAN) art generators offer the imagination an entirely new space to play.
GANs are where machine learning thrives. Coined by researcher Ian Goodfellow in 2014, GANs use only two neural networks to achieve generative art. One neural network—generative—trains the algorithm using a dataset to then generate a similar image. So, for example, if you wanted to create a photo of a dog, the algorithm (AI photo generator) would learn from millions of other pictures of dogs until it can construct an image of a dog itself.
The second neural network—adversarial or discriminator—takes the new images generated by the first neural network to distinguish what images have been produced by humans and what images were produced entirely by the algorithm. This process continues until the algorithm is fooled.
While the output may mimic similar styles to the input, it becomes an entirely new image instead of a manipulated image.
And the best part? No coding experience required!
How to Make AI Art
Want to make AI generated art yourself?
If you have an idea for an image–or even a whole series–that you'd like to turn into something real and tangible, there's no need to spend hours in Adobe Illustrator. Instead, with the help of AI art generators, just plug in some parameters and let the generator do its thing!
When working with AI image generation, it's important to create effective prompts that assist the AI model in learning to generate images that are accurate to your request.
The more specific and in-depth a prompt is, the more control you'll have over the output. If you're not sure how to get started in creating AI art prompts, here are a few ideas to get you started:
- Consider the style you're looking for. Do you want a surreal look and feel? Maybe you're a fan of the impressionist style. Referencing art styles is a great way to ensure the output is exactly what you want.
- Subject and details. Make sure you include a specific subject–such as building, plant, or woman–and what details of the subject you want to ensure are included. What kind of surroundings might they have? Do you want a specific color included?
- Use a minimum of 5-7 words. Using only a few words leaves room for vagueness in your prompt. Make sure you are using multiple references and description words. For example: "dog playing ball" leaves a lot of details up for interpretation, but "fluffy dog playing with large ball, surrealistic" might give you exactly what you need.
- Use ethical prompt references. While some generators filter inappropriate words and artist references, not all do. AI should not be used to produce inappropriate or offensive images, nor should it be used to rip off the works of previous artists. Avoid referencing specific artists or using NSFW language.
Some platforms, like Lexica or LensAI, don't even require text prompts, and will generate multiple options for you. For the apps taking over social media feeds, all you need to do is upload a few selfies for them to develop a wide variety of portraits.
AI Generated Art Elevates the Creator Economy
Beyond transforming your headshot into a Victorian portrait of a princess, what can we do with this powerful technology? The mind's the limit!
AI art generators from text will be an excellent tool for visual creatives who want to experiment with different styles without investing in expensive software packages or learning how to code. Instead, they can use these tools to create unique artwork from scratch or remix existing works of art using the algorithm's suggestions.
These AI image generators can also be used in the design process by providing inspiration for new designs and ideas, which can save time and money when creating new work for clients.
Thus far, we've seen AI-generated art used in advertising, architecture, fashion, and film. Historians can generate historically accurate images of civilizations' past. Comic book writers can produce volumes of stories like never before. So far, we've seen creatives generate the following:
- Human faces
- Cartoon and anime characters
- Basic sketches
- Floor plans
- Animated series
The list goes on!
But don't take our word for it. Let's take a look at some real-life creators embracing AI.
Creators Doing it Right
AI is starting to be viewed as a business technology solution.
Creators are using AI to revolutionize their industries. AI art generators help bloggers, designers, and many more with their creative projects. Anyone in the process of switching careers to a more creative profession now has a faster on-ramp. Innovation in these industries is continuously growing to a new level, thanks to innovative minds using this AI software to bring something entirely unique to their craft. Here's how they're doing it.
AI Generated Content
The best kind of storytelling is visual storytelling. As technical content creators and storytellers seek to push their creativity forward, AI could be the answer.
With AI's power, agencies can produce volumes of graphics for social media clients. Independent creators can produce videos in minutes without having to learn Adobe Premiere. Journalists can recreate scenes without the use of a camera.
Take the LA Times, for example. In one of their latest projects, AI Sees LA, they used the leading AI art generator, Midjourney, to create surreal and stunning images representing the multi-faceted city. The result: a visual poem using literature, mythology, and prose.
Digital content creator, Fabio Comparelli, started in landscape photography and has recently embraced the world of AI-generated art. One of his most viral pieces is a video visualizing human evolution, which was generated with Midjourney.
AI art makers democratize content creation, and as the technology becomes more visual-forward, it will serve as a must-have tool for creators in the future.
AI Generated Films
As we can see with Fabio's work, AI art generation isn't limited to static images. So there's a lot of promise in AI for film, too.
Fabian Stelzer (not to be confused with Fabio) has used images generated by Stable Diffusion, Midjourney, and DALL-E 2 to produce a science-fiction short film series titled "Salt". He's also using AI voice generation for the voiceover and a text generator to assist with script writing.
AI Generated Fine Art
"Even historical figures like Michelangelo outsourced labor to create iconic works like the Sistine Chapel. AI offers artists a similar ability to focus more on theory, meaning, and abstraction, rather than execution, which in many cases takes years of painstaking labor to realize a single piece." - I Make Art With A.I. Here's Why All Artists Need to Stop Worrying and Embrace the Technology, by Agnieszka Pilat.
A painter by trade, Agnieszka has had a longstanding fascination with technology, as machines are her main subject in her work. In her most recent show, ROBOTa, Agnieszka will showcase paintings completed by AI and an ensemble of robots in San Francisco.
AI Generated Video Games
Co-founded by Emmanuel de Maistre, Scenario is a platform that enables game developers and artists to create high-quality video game assets with Stable Disffusion's AI. From creating dreamy characters to vast other-worldly landscapes, video game creators are empowered to focus on the details of the game while leaving the bulk of art direction up to tech.
Meet the AI Art Image Generators
While AI art generation is destined to democratize creativity, it also presents endless opportunities in the tech industry. Within tech, the AI art industry is comprised of a few different types of companies:
- Those who produce language models
- AI generator startups
- Companies incorporating AI generators into existing platforms
Let’s take a look at the key players in each category and how they’re making AI art their own.
Main Language Models
Language models serve as the foundational element of any AI art generator. In the tech world, language models are defined as "a probability distribution over words or word sequences". In lay terms, a language model provides the probability of a specific sequence of words being valid. The validity or invalidity of a prompt helps in machine learning. Therefore, the AI algorithm learns what can and cannot be a valid output based on the prompt.
Sounds simple right? Not so much. That's why much of the AI art generator industry is looking to three main language models to democratize this technology:
- Stable Diffusion
- Open AI’s DALL-E and DALL-E-2
- Niji Journey
You've already heard us talk about Stable Diffusion's incredible impact on the AI art industry, but we'll provide a little more context here.
Stable Diffusion is an open-source AI generative model created by Stability AI that enables text-to-image art generation. This means anyone on the internet can use the language model to create text-to-image tools for creators. Since the language model was released in August of 2022, the company has raised $101 million in funding, released Version 2, and had over 200,000 downloads as of October.
AI research startup OpenAI created DALL-E and became one of the earliest players in AI art generation in January 2021. The goal of DALL-E is to demonstrate the potential of AI to create novel and unique images based on natural language descriptions.
In September, DALL-E 2 was released. DALL-E uses the GPT-3 language model to create original, realistic images from a text description.
Unlike Stable Diffusion, DALL-E is not open source per se. It’s free to use, but there is a limitation on how many free credits you get. Users are allotted 50 credits during their first month and 15 free credits after that. For anything additional, you’ll have to pay up.
As an independent research lab, Midjourney has also produced an AI platform that creates images from text. Known for its highly artistic style, Midjourney isn’t the go-to for more realistic images, but many artists and video game creators are flocking to it for its painting-like output.
Similar to DALL-E, Midjourney isn’t entirely open source. Some features can be used for free to a limited degree. Then, there are tiered subscription plans based on how much you intend to use the technology.
Niji Journey is an AI that draws custom animate illustrations. The Niji model is based on the Midjourney model, rained to produce anime and manga-style artwork. The model was created with Spellbrush (creators of Waifulabs, one of the earliest AI anime generators) and Midjourney.
Ready to make a new anime series? You can join the waitlist here.
The New AI Art Companies: Building from Zero
Thanks to the rising popularity of AI, there has been a wave of new AI art startups applying artificial intelligence technologies to create digital art. From animations to avatars, there’s something for everyone.
Some use neural networks to generate paintings based on photographs or video, while others use computer vision to create new materials based on existing ones. In fact, there are already companies that are employing both techniques.
Let’s look at some of the most prominent companies in this space!
List of AI-Image Generator Companies
Unlike design software, image generators have one function: to generate unique images based on a text prompt or pre-existing materials.
- NightCafe - Users can create art and order prints with text-to-image or style transfer creation methods.
- Jasper - A content generation platform that helps users generate social media posts, ad copy, blog articles, emails, website copy, art, and more.
- Craiyon - Formerly known as DALL-E mini, users can generate an image with a text prompt and print it on a shirt!
- Fotor - This AI-powered photo editor helps users generate images and art or transform existing photos.
- Photosonic - This free text-to-image generator lets you create images in various styles, including Anime, 3D, and cartoon.
- Deep Dream Generator - Users can generate art and photorealistic images from a text prompt with or without a base image, merge a base image with a famous style of painting, or use a deep neural network to generate psychedelic abstract art.
- Rosebud.AI - Users can access AI-generated stock photos and video, generate a realistic image with text, or animate portraits.
- Playground - As a free-to-use AI image creator, users can create art, social media posts, presentations, logos, and more.
As one of Lexica's early investors in 2020, Worklife has a soft spot for the platform. Unlike the design platforms, Lexica serves as a massive database of images generated by Stable Diffusion AI.
- Search millions of AI-generated images
- Free AI image generator
- Remix images you discover
Once you select an image, you can see the AI art generator text prompt, explore images in a similar style, and generate new images using the original image. As an open-source platform, these images can be used by anyone for any use.
List of AI-Avatar Generator Companies
Face recognition is one of the major benefits to AI and some companies are leveraging the technology to build avatar generators. (Yes, this is what’s taking over everyone’s social feeds!)
Some of the top AI avatar generators on the scene include:
- Lensa - This all-in-one image editing app just rolled out a “magic avatars” feature that allows users to transform a selfie into various styles of images.
- Avatar AI - For just $18, users can create over 100 avatars using a collection of selfie images.
- Artbreeder - In addition to regular image generation, users can remix images and edit facial features for a totally custom portrait.
List of AI-Generated Video & Motion Companies
AI has promise beyond static images and art. We’re also seeing a rise in AI-powered video and motion graphics platforms! Here are a few of our favorites:
- Runway - This AI video editor provides tools that enable users to remove video backgrounds, paint out objects, automate motion tracking, and more!
- Synthesia - Create realistic videos from text prompts in minutes!
- Wombo - Wonder where those lip syncing videos of famous people come from? Probably Wombo. Take a selfie or use a headshot, pick a song, and let it sync!
- Tavus - Transform one video into hundreds of personalized versions with this programmatic video tool. You can swap out variables to change the script with just a few prompts.
- Rephrase.AI - This platform generates personalized videos with digital avatars to help you scale your video creation and engagement.
- Hour One - Generate a custom explainer video with text using your own digital avatar.
- Luma AI - Transform a video into a lifelike 3D interactive model.
Companies Incorporating AI Art Image Generators
More companies are embracing the idea of art made by AI, and it's becoming more mainstream by the day. As a result, we're seeing many companies adopt this technology and integrate it into their products to make them more efficient.
For example, Adobe's new AI-powered Photoshop tool allows users to edit photos on their phones, or Microsoft's Windows 10, which uses AI to help you find files faster.
As noted in the chart above, many companies are rolling out their AI art generation algorithms. Many others are incorporating their technology into their software and online apps. Stable Diffusion's open-source AI model makes it a top pick for integration into creative platforms. We're going to showcase the platforms that are putting this incredible technology to work, including:
- Adobe Creative Suite
Need help telling a compelling story visually? It’s never been easier with Tome (Worklife-backed company), a storytelling tool that gives your work the edge it deserves. With frictionless drag-and-drop, Tome can help you tell your idea with any type of content: embeds, 3D renderings, Figma embeds.
Tome is now using DALL-E 2 to bring your story to live. See it yourself:
Adobe Creative Suite
Storyteller, technologist, and current VP of Product at Stability AI, Christian Cantrell, moved quickly to create a plugin for use in Adobe's Photoshop. Users with a DreamStudio account can install the plugin and generate images using the text-to-image generator. In addition, the tool has also enabled an image-to-image option that transforms pre-existing pictures.
A popular Photoshop plug-in is Alpaca, a new Photoshop plugin that allows artists to change and enhance their pictures. The tool allows you to: inpaint, transfer, and concept generation, as well as the ability to upscale images.
Canva has already proved to be a significant disruption in the design world by itself. In 2013, Canva was launched to democratize design with an intuitive, drag-and-drop design platform. Fast-forward to September 2022, Canva took things a step further when they released a Text to Image app powered by Stable Diffusion AI. Canva Pro users can use this free tool to generate an image with only a text prompt.
Figma is one of the leading design platforms, making a name for themselves against the giants Adobe XD and Sketch. Like Adobe, plugin creators are flocking to Stable Diffusion to enable text-to-image generation in Figma.
One leading the charge is Antonio Cao's plugin, designed to use the power of technology to turn simple shapes into a fully fleshed-out design.
Diagram's Magician is another Figma plug-in that allows you to generate unique icons and graphics from text.
Remember: Embrace Ethical, Responsible AI
As AI image generators become increasingly popular, the technology is quickly being adopted by artists and creatives. But it's important for users to ensure their work is ethically created. This means ensuring their content is ethical, responsible, and meets all legal requirements.
So what are the main ethical concerns spotlighted by creative leaders like Jon Lam in his recent piece with NBC? Much of the controversy centers around using pre-existing works of art to feed machine learning without permission from the original artist.
This brings about two main concerns:
- Is the generated art considered a copyright infringement if it's "stealing" a portion of an original piece of art?
- Does AI-generated art reduce the value of handmade art enough to replace artists?
While we're still in the early phases of AI art adoption, precedent tells us that AI and art will find a balance that satisfies everyone.
Prisma Labs is one of many platforms refuting the claims.
The company goes on to say that the art generated cannot be considered exact replicas of any single piece of art, because the algorithm ensures the output is entirely unique. In addition, these platforms include terms and conditions that allow the technology to use any images generated as additional training for its AI model.
Another example of technology companies doing their best to answer these questions is Stable Diffusion's latest update. Stable Diffusion Version 2 includes new restrictions prohibiting prompts from mimicking other artists' work. It's a big step in the right direction toward protecting the value of art and the intellectual property of artists. In addition, an adult content filter was added to limit the generation of NSFW images.
So what about AI-generated art replacing the work of artists?
When Napster started making music available for free, artists relied almost entirely on album sales. It stripped away their primary source of income overnight. Now, revenue in the music industry has a completely different model, even though Napster is long gone.
Digital streaming stuck around with platforms like Spotify and Apple Music, and performers found new ways to make money. Today, their main revenue streams include ticket sales, merch, music licensing, and the rise of celebrity-owned brands.
The tech industry expects the same to happen with AI-generated art. There will always be an important place in the world for hand-crafted art. That will never change. In addition, we expect to see a wave of new artists hitting the scene and seasoned artists testing their creativity in a brand-new medium.
At Worklife, we believe in the power of AI tools that can help anyone become a thriving creative. And we look forward to seeing all of the incredible things we can create with it. From content generators like Copy.ai to Canva's text-to-image plugin, we expect to see high-quality content, art, film, and video games produced at an unprecedented rate.
And AI won't stop there…
The future is here, and we're ready to live in a world where workers are empowered by AI to be more precise and productive.
Want to dip your toes into the AI world? You can learn more about the movement to keep AI ethical at MIT's RAISE. Here are a few other really awesome tools we suggest you check out:
- ChatGPT - A chat bot that leverages AI technology, enabling users to have human-like conversations, compose essays, generate descriptions of art, and even code!
- AI For Everyone - This Coursera course is taught by Andrew Ng, founder of DeepLearning.AI and provides a basic understanding of AI for non-technical people.
- The Creativity of Code: How AI is learning to write, paint, and think - Written by author and British mathematician, Marcus du Sautoy, this book explores the nature of creativity and an overview of hour AI algorithms are replicating creativity to produce all kinds of creative work.
- Drip.art - Described as “your personal AI artist”, you can get a free AI generated portrait of yourself or subscribe to generate whatever you want!
- Sketch2Code - A Microsoft AI Lab tool that converts hand-drawn wireframes to working HTML.
P.S. AI avatars are everywhere? Want to learn more? Read more below:
- Lensa AI Transforms Your Photo Into an Array of Movie-Like Avatars
- What You Should Know Before Using the Lensa AI App
- (Spanish) Así funciona Lensa, la viral aplicación que transforma tus selfies en avatares digitales
Did you enjoy this article? You might also like: