In the world of AI, one of the most captivating advancements is in the realm of visual art – Generative AI has enabled the creation of AI-generated images or AI art, essentially turning text into visually striking images.
Delving into behind-the-scenes stats and figures is one of the best ways to fully understand generative AI's transformative power in the creative and visual industries.
From how many AI images exist to how pervasive AI-generated images have become, what are the most popular AI tools, and how much money they are worth, this visual AI stats overview unveils it all.
Strap in to unearth the real impact of visual AI in under a year since its inception!
|Total number of AI Images created||15 billion|
|Daily AI Images created||34 million/day|
|Biggest source of AI-generated images||Stable Diffusion – 12.59 billion images|
|Fastest growing AI image generator||Adobe Firefly|
|AI market value (2022)||$136.55 billion|
|Projected AI market value (2030)||$2 trillion|
|Highest-valued AI company||OpenAI – $20 billion|
|People who can recognize AI Art||54%|
|People who enjoy AI visuals||56%|
Before we get down to the numbers, let's quickly review what those numbers are quantifying.
Generative AI is artificial intelligence capable of generating content such as text, images, or other media from user's instructions. It employs generative models that analyze data from provided training sets to identify patterns and structures and later replicate them in entirely new data.
AI-generated images or AI art pieces are digital, synthetic pictures created with an AI-generative model. They're different from “traditional” pictures that were shot with a camera or painted digitally or physically by a human artist.
The most popular type of AI image generators are those that use text prompts –descriptive, written instructions of the images to be created– input by human users. They're called text-to-image generators because they essentially transform text into images.
Let's get started by overviewing just how big this novel visual AI space is already. A simple way to do that is by looking at how many pictures this technology has produced to date.
Our friends at Everypixel have a great research report that estimates that right now, in the world…
There are 15 Billion AI images already
To put this in perspective, this huge number of images was created in the one year since the first publicly accessible AI image generator was released in September 2022. In contrast, it took humans 150 years to produce that amount of pictures.
For comparison, these metrics help:
There are a few considerations that these stats are not accounting for yet, however. We wonder –and are working on getting the answers–:
How many of these images are usable for their intended purposes? How many are even visually accurate? In our humble experience, it can take a while to get the gist of text prompt writing and more than a few attempts to obtain a decent result, especially if you're not photography or digital illustration savvy.
Courtesy of Everypixel, we know that…
There is an average of 34 million AI image creations per day
Considering how recently this technology has reached the public and how relatively small the offer in tools is still, this is another quite staggering cipher.
As the trend in text-to-image generators is to upgrade and evolve, at a vertiginous fast pace as well, to produce better visual results from simpler prompts or by adding AI language models to assist users in text prompt writing or a combination of both, we can only expect that such improvement impacts the number of synthesized images per day, exponentially, in the coming months and years.
Next, let’s have a look at the most used generative AI tools for image creation and their particular metrics to understand better where are all these images coming from.
|AI Image Generator||Total Images Created||Notable Points|
|Dall-E||916 million||The most popular AI image generator|
|Stable Diffusion||12.59 billion (80% of all AI images)||Fastest-growing AI image generator|
|Midjourney||964 million||Most visited website|
|Adobe Firefly||2 billion (in 6 months)||Fastest growing AI image generator|
|Shutterstock Generate||N/A||Launched before Adobe Firefly, tapping into 550,000+ subscribers|
|Getty Images||N/A||Recently launched|
|123RF||N/A||Incorporated AI generator in July 2023|
Dall-E has the honor of being the very first AI text-to-image generator released to the general public. Its first version was announced in January 2021; Dall-E 2 rolled out for regular users between August and September 22, and finally, Dall-E 3 is being gradually released in September 2023.
In this time, the tool has amassed around 1.5 million users who produce an average of 2 million new images daily, making this app contribute 916 million files to the pool of AI-generated images.
Did you know? OpenAI is currently the highest-valued AI-focused lab, with a valuation of $20 Billion, but despite being a pioneer in this space, Dall-E remains the lesser-used of all the main AI image generators.
Midjourney is a popular AI image generator released in July 2022 by a self-funded and independent research lab. In a little over a year of life, it has reached 17.9 million users, who are said to have created around 964 million images so far, breaking down to about 2.5 million images processed in a day. A lot of users highlight that it produces higher-quality results than other tools, particularly in terms of photorealistic imagery.
Fun fact: It has the most visited official website, with over 35.5 million visitors, despite the fact the generative model is available via Discord only (not on the website), where all its users must register to use the tool.
Stable Diffusion is a generative AI model for images developed by the tech lab Stability AI, originally launched in August 2022. The company has the model available on two different host platforms: Dreamstudio and Hugging Face. Across both channels, it's said to have a total of 10 million users and to be producing about 2 million images a day.
However, Stable Diffusion is an open-source model and is available to third-party developers. It's nearly impossible to calculate how many tools are built on this model and how many users each of those has. But a rough estimate by Everypixel suggests there are tens of thousands of Stable Diffusion-based models and 11 billion+ images created thus far across the most popular of those third-party apps.
This puts the total of images generated by Stable Diffusion at 12.590 billion.
So, guess what? Stable Diffusion is the most popular AI image generator of all, responsible for 80% of all AI-generated images today. Curiously, and due to its open-source nature, it has the less-visited website among competitors.
Not-so-cool data: It has been sued separately by Getty Images and a collective of three artists for copyright infringement on its training datasets.
Media licensing specialist companies have worked and continue to work in developing their own generative AI tools for visual arts. Here are the main ones and what we know about their performance.
Adobe’s proprietary generative AI model for image editing and image generation, Adobe Firefly –trained only with authorized Adobe Stock files and non-copyrighted content– came out much later than independent tools. The first release, in beta, was in March 2023. However, it saw exponential adoption and growth: in less than six weeks, over 100 million images had been created with this text-to-image generator, and after it began being integrated into Adobe Photoshop and Adobe Express, two of the platform’s most popular apps, the numbers exploded. In May 2023, they had reached 1 billion image generations, and in September 2023, on the full public release of the models and only 6 months after its first launch, they announced the 2 billion images milestone had been reached.
Numbers say it: Adobe Firefly is the fastest-growing AI image-generating tool in terms of popularity and user adoption, and we bet it hugely benefits from being directly exposed to Adobe Creative Cloud’s 30 million users.
The stock photo titan Shutterstock released its native AI generative feature, Shutterstock Generate, in January 2023. This service is built on the base of the Dall-E model but trained exclusively with Shutterstock content. In May 2023, they announced an upcoming feature for this tool, a virtual design assistant that performs image generation and image edits in a chatbot-conversation style.
The company hasn’t shared any relevant metrics for how many customers use the feature or how many images it has generated so far. But we can highlight that it was launched a couple of months earlier than Adobe Firefly and that the firm aimed at tapping into its 550,000+ subscribers base to adopt the tool.
Getty Images is one of the latest media licensing firms to join the AI industry with the recent launch of Generative AI, a text-to-image generator powered by the NVIDIA Edify model and trained solely on Getty Images content, including premium and exclusive imagery.
As the feature has just been released in late September 2023, there are not a lot of significant numbers yet.
The well-known stock photo site 123RF incorporated its very own text-to-image AI generator back in July 2023. While details on the model or the training dataset were not disclosed, the company did share that the images generated were under their custom license agreement. Later on, they also introduced AI image variations to see four new variations of every stock image in their library. We don’t have more stats from them as of yet.
Let’s see some of the most important market stats for visual AI to see what is worth what.
|AI Market Stats||Value|
|Current Market Worth||$136.55 billion|
|Projected Market Worth (2030)||$2 trillion|
|Largest Regional Market Share: North America||36.8%|
Per this report from GrandViewResearch, we know that…
The AI Market is currently worth $136.55 billion
If that made you dizzy, wait until you see the projections for the end of the decade:
It's expected to be worth $2 Trillion by 2030
In case you were wondering, we did the math: that is a 13x growth in under 10 years.
The stat we are still missing but working on defining is how much of that market will be generative AI and what slice of that pie will be visual AI.
While that number is still not specified, we know for sure that image generators, chatbots, and mobile apps are largely responsible for fueling that growth rate.
In the current scenario, it shocks no one that…
North America is the regional AI leader, with 36.8% total AI market share in 2022.
This represents $736 Million. A market study by Global Industry Analysts (GIA) states that the US AI market share will be close to $300 billion by 2026.
Projections made by experts at Facts and Factors also say that
China will catch up and reach 26.1% market share by 2030.
So, unsurprisingly, the two largest global economies will dominate the AI space before the end of the decade.
Interestingly, most researchers point out that in North America, the dominant segment is generative media –including visual AI–, whereas the growth of the AI market in Asia will be mostly linked to LLMs (Large Language Models such as AI chatbots and text or voice-based virtual assistants)
Which are the biggest –most valued– players in this novel industry? Here:
And this only accounts for AI tech labs focused on AI products. You also have to consider all the non-AI specialized companies entering the AI race, particularly in the last year: Corporations like Adobe, Amazon, NVIDIA, and Shutterstock, with high market valuations, that develop and release their proprietary generative models.
|Americans who have seen AI Art||27%|
|People who can recognize AI Art||54%|
|People who enjoy AI visuals||56%|
|Preference for AI Art||34%|
|Preference for Human Art||29%|
|AI Art and Human Art perceived as similar||31%|
The last statistics we want to look at to wrap up this overview of visual AI are those reflecting the user perspective: people who view AI art and people who create it.
An insightful report from YouGov collected data that says…
27% of Americans saw AI art in some form
Be this a representation of the global experience, one has to consider that despite the current frenzy in the creative world about AI-generated art, it's still far from taking over mass visual production. Yet this is a very high percentage of exposure for a type of art that essentially didn't exist on the Internet 15 months ago.
Do people like AI-generated images? This is an important question that sometimes gets overlooked. The same survey threw that…
56% of people who saw AI visuals say they enjoyed it
So, it seems audiences will continue to fuel the rapid adoption of visual AI.
Finally, the most controversial of the socially-oriented polls in generative AI media. Stats say that when asked to compare AI-generated visuals to human-made ones…
Of course, these results are based on subjective answers from a non-expert pool of respondents. But with that in mind, it's rather interesting how tied the results are.
Experts at Yale surveyed 500+ undergraduates for a report and concluded that…
54% of surveyed individuals could tell AI art from traditional human-made work
We are not shocked at this answer. At this point, there is often a “synthetic quality” present even in the most convincing AI-generated pictures that give away they're not man-made. An artistic opinion could be that AI images still lack a “creative soul.” But it's also true that when created by the average, non-art expert Internaut, AI visuals are not yet as photorealistic or artistically as valuable overall.
There are clear exceptions, though.
Fun fact: In 2023, the renowned German photographer Boris Eldagsen won the prestigious Sony World Photography Award… with an AI-generated image he crafted. Eldagsen would go on to decline the award for ethical reasons, but his image fooling expert judges to this degree certainly sparked a debate on the quality and potential of AI imagery.
This concludes our quick look at the latest available AI stats for generative media.
In our opinion, the ciphers for this one-year-old space are huge enough to impress anyone and do nothing but cement that visual AI is here to stay and, at this rate, set to take over the creative industry.
However, there is more than one key metric still not factored in; many breakdown numbers are missing that could help really measure how impactful AI visuals are in the world and what size the generative imagery market can really reach.
Like with everything in this novel industry, it's a matter of waiting and seeing.
What do you think about these AI stats? Feel free to share in a comment below!
I am an experienced author with expertise in digital communication, stock media, design, and creative tools. I have closely followed and reported on AI developments in this field since its early days. I have gained valuable industry insight through my work with leading digital media professionals since 2014.
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