In just one year, a survey found that over 15 billion AI images were made. That’s more than the total number of photos taken in the first 150 years of photography history. It’s pretty impressive when you think about it.
AI-driven artistic communities have exploded online over the past year. From Reddit to Twitter to Discord, there are tons of places where AI “artists” gather to improve their skills and show off their creations to the world.
The amount of content generated during this time is hard to quantify, but it’s truly remarkable. The statistics from this study on AI-generated images show a massive shift in the art world. It raises questions about creativity, technology, and the future of visual expression.
One of the most astonishing findings from the report by Everypixel is the sheer number of images produced by text-to-image algorithms in just one year. To put it in perspective, it took photographers 150 years, from 1826 to 1975, to reach the mark of 15 billion images. It’s mind-boggling to see how rapidly things are expanding.
Some major players in the field are OpenAI’s DALLE-2 and Adobe Firefly. DALLE-2, released in April 2022, has been generating an average of 34 million images every day. Adobe Firefly, integrated into Adobe Photoshop, created an incredible 1 billion images within just three months of its launch.
Midjourney, a popular AI model, has also made significant contributions to this digital art revolution. With 15 million registered users and 1.5-2.5 million active members, Midjourney’s Discord channel alone sees a constant stream of image generation. Estimates suggest that up to 2.5 million images are produced daily.
Stable Diffusion, an open-source text-to-image model offered by Stability AI, has also gained traction. With over 10 million users and roughly 2 million images generated daily, it has made an impact. However, due to the model’s open-source nature, accurate statistics are hard to come by, as many derivative platforms have emerged.
Collectively, these pioneers have generated over 11 billion images on popular platforms like GitHub, HuggingFace, and Civitai. Taking other models into account, such as Runway and Stability AI’s official channels, the total count using Stable Diffusion technology amounts to an impressive 12.590 billion images, making up 80% of all AI-generated images.
To put the scale of these 15 billion images into perspective, it surpasses Shutterstock’s extensive library and would represent one-third of all images uploaded to Instagram.
Looking back at the launch of Midjourney in June 2022, I am amazed at how much the quality of AI-generated images has improved in just a year. It’s nearly impossible to distinguish between a photo and an AI-generated image now. It’s exciting and terrifying to think about where AI will take us in another year. The creative industries are transforming right in front of us, and the possibilities are endless. So buckle up and enjoy the ride.