AI-generated art has been around for several years, although it wasn’t until recently that it became mainstream. In terms of its market, AI art creation platforms fall into the AI-generated content (AIGC) bucket.
DALL-E 2 is one type of AIGC. It creates images of comparable quality to those produced by human artists by inputting certain keywords into the machines’ parameters. And that ease of use translates into big business. According to InsightSLICE, the global AIGC market will reach $38.2 billion by 2030.
In early 2022, a series of AI-powered algorithms were introduced that gave users of all skill levels the power to generate images from single-line text. And these programs have steadily grown in number and complexity, reflecting an insatiable user demand for a diverse range of artistic styles and content types.
For example, NightCafe Studio users have generated more than 75 million images and DALL-E 2’s 1.5 million users are generating more than 2 million images daily. However, what those numbers don’t reveal is where the industry is headed.
As an industry insider, I can assure you that it will look a whole lot different than where it is today.
Before looking ahead, it’s worth looking back on how far the industry has come in just a few short years. When I first entered the industry, there were only four or five AI art generators. This was in 2019 before text-to-image algorithms existed. Back then, the only AI art algorithms were “deep dream,” which made images look like they had dog noses all over them, and “neural style transfer,” which copied the style of one image and transferred it to another.
When text-to-image algorithms came out in early 2021, starting with the announcement of DALL-E followed by some open-source algorithms like VQGAN+CLIP, a few more AI art generators popped up. But by the end of 2021, there were probably only another ten or so apps. Then, In 2022, the space started accelerating and finally exploded when Stability AI released “Stable Diffusion” as open source.
As of now, there are far more AI art generators than even I can keep track of. It’s easy enough now to create a new AI art generator app as a side project, so there are hundreds of those. But we’re also seeing big players like Microsoft and Canva, and slightly smaller ones like Picsart, and other big mobile apps add text-to-image generation into their existing product offerings.
AIGC, and more specifically, AI image generation as an industry, is still extremely young. The big players are generally less than a year old, and the big names that have adopted image generation into their services have only done so within the last month or so. But today, we’re nearing the point where AI will be good enough to use in video games. Concept artists and traditional artists are already using AI image generation for inspiration, and even incorporating the results into their artworks.
However, in the end, AI-generated art could be a victim of its own success. Progress on image generation technology is moving so quickly that it’s nearing a point where outputs are nearly perfect. And when this happens, which will be soon, the algorithms and models will become commoditized and the winners will be defined less by their AI models and more by the technology and user experiences they build around those models. That means the industry, in its current form, could cease to exist.
AI image generators may simply be a function of larger platforms and applications. There might still be some powerful, specialized AI-powered image and/or video generation apps for pros, or maybe they’ll all end up as features of existing tools like Photoshop and Final Cut. Canva and Jasper are already diving into the space, and I can see a day when Salesforce, Hubspot and even proposal software like PandaDoc could include AI-generated art creation.
And once image generation becomes “solved,” the smart people working on the technology will move on to other modalities like audio, video, writing or other problems.
Even after AIGC becomes intertwined with systems like CRMs, marketing automation platforms, Web development tools and more, the AI art hobbyists — who just create art with AI for fun or self-therapy — will still exist.
Creators will also continue to monetize their time by utilizing AI-generated art platforms by selling their work as NFT collections, prints for print-on-demand service, stock photography or even on talent platforms like Fiverr and Upwork. Others will continue to use AI image generation to save money on things like book or album covers or in place of stock photos on a blog article.
But, the most overlooked use-case is people simply using AI art generators for fun or to wind down. Many people love creating art with AI. We often get feedback that AI art has scratched a creative itch that people didn’t know they had. We also often hear that people use it religiously as a way to wind down and de-stress or forget about the outside world for a while. For some subset of the population, AI art triggers a dopamine hit and a sense of pride and accomplishment that people simply couldn’t get before.
For many of us in the industry, the problem we are solving is how to best democratize art and other creative services. And I believe we have made progress. Since discovering AI art, it’s become many people’s number one hobby and they do it every day. Before AI, learning to create beautiful art took thousands of hours of practice.
With AI, the same sense of pride and accomplishment can be gained in just a few minutes. Whether they are new users typing a few words and choosing a preset style, or proficient users adjusting a wide range of settings and trying new algorithms, people are proud of their work. They discuss it online, share prompts and tips, complement each other’s work and share their favorite images.
I don’t think that’s going to change; an entire industry will develop around these “hobbyists.” I’m betting my livelihood on it!
(Copyright: VentureBeat What you need to know about AI-generated art (venturebeat.com)
Chinese
Japanese
Korean
