Where To Sell AI Art? – Part 1. Stock Image Sites

In this newsletter, read about:

• 🕵️‍♀️ Stock Image Sites & AI Art

• 🗞 News and Top Reads

• 🏪 AI Art Tutorial: Hyperrealistic Midjourney Images

• 🎨 Artist of the Week: Refik Anadol

• 🖼 AI-Assisted Artwork of the Week: Majesty of the Night Sky

• 🤓 How to Get Started with Generative AI?

🕵️‍♀️ Stock Image Sites & AI Art

If you have spent some time with AI image-generation tools, you probably already know that AI art can actually be extremely time-consuming, especially when you strive for perfection. So, one day or another, you start to wonder – may I sell these images?

To help you answer this question, I’ve researched how the largest stock image sites deal with AI-generated content. Here’s what I’ve got.

Shutterstock

Shutterstock has taken a very interesting approach to AI-generated content, which probably reflects that this platform focuses on visual content rather than art. So, they provided their huge image database to OpenAI & LG to train AI-generating tools. Now, they suggest that customers generate images using Shutterstock’s own AI-generating platform – and then, pay for these images, while artists who provided their content for training the corresponding models, we’ll be compensated.

Yes, you create an image and you pay for it. Okay, let’s say it can be fine because, after all, DALL-E 2 and Midjourney also charge you for using their AI models and computer resources. But there is a huge difference. You own the images you generate with Midjourney and DALL-E 2 as a paid customer. And when you generate with Shutterstock, the image is owned by Shutterstock:

So, you create a beautiful image using Shutterstock AI Image Generator, trained on Shutterstock images. You pay to download this image. Then, this image can be offered to other customers, and the amount they pay will effectively go to Shutterstock and all other artists on the platform, whose images were used to train the model. But not to you – sorry 😂

Still, note that compensation to original artists is likely to be very insignificant, considering the huge amount of images used for training the model (i.e., there are 400M+ images in the Shutterstock library). But it’s interesting that they consider the authors of the original content to be the ultimate authors of the AI-generated images, while people who generate these new images are not only refused the role of an author or an artist but need to pay for the images they created, and are not compensated when these images are sold to other customers.

I was very surprised by this discovery, knowing how much effort AI artists put into engineering the right prompts, experimenting for many hours to get the desired outcome, and often using Photoshop to make the image perfect. But then, I realized that it’s probably very different for Shuttestock’s customers, who are primarily not looking for artwork but usually something simpler to put into a PowerPoint presentation or a blog post.

Still, I cannot fully agree with this approach because AI-generating tools are not perfect at their current stage of development, and you still need many attempts before getting to the desired result – if you ever get there! So, Shutterstock’s model is definitely not my favorite.

And if you had any doubts, Shutterstock doesn't allow uploading images generated with other AI tools, to their platform:

They present this as a responsible and ethical approach to AI-generated content but I think they tend to overestimate the role of the individual training images and underestimate the role of prompt engineering. As text-to-image AI models get better, Shutterstock’s approach might work for generating some basic visual content, but to generate artwork, you still need a human guiding the AI tools, and Shuttestock is on its way to losing this part of the market.

Getty Images & iStock

Similarly to Shutterstock, Getty Images decided to ban AI-generated content.

Moreover, in January 2023, Getty Images started “legal proceedings in the High Court of Justice in London against Stability AI claiming Stability AI infringed intellectual property rights including copyright in content owned or represented by Getty Images.”

In contrast to other AI image tools, Stable Diffusion’s training dataset is open-source, and apparently, images from stock sites, like Getty Images, constitute a large portion of its contents. This is also evident from Getty Images watermarks found on some Stable Diffusion images.

It might look like Getty Images is at full-scale war against AI-generated content, but this is actually not true. Getty Images CEO recognizes that AI image tools are “just the latest example of technology expanding the amount of available imagery”, like digital cameras, smartphones, and social media before them.

It doesn’t look smart to simply ban technological advances. So why would Getty Images do this? Actually, they don’t. In their statement about legal action against Stable Diffusion for improper use of copyrighted images, they also mention that “Getty Images provided licenses to leading technology innovators for purposes related to training artificial intelligence systems in a manner that respects personal and intellectual property rights.” So, it’s very likely they will follow the Shutterstock example, and will only allow AI images generated with their own tool.

Adobe Stock

This platform can be actually considered a pleasant exception for AI artists. Adobe Stock allows AI-generated content as long as it meets their “quality, legal, and technical standards”.

To submit an AI-generated image on Adobe Stock, you need to:

• ensure that terms and conditions for your generative AI tools allow commercial use of generated images;

• don’t submit works depicting real places, identifiable property, or notable people;

• label your image as a Generative AI illustration;

• upscale an image so that it meets existing visual and technical quality standards for illustrations.

I find this approach very smart and honest. First of all, they are getting their share of the AI art market, which can grow pretty huge in the coming years.

Secondly, they don’t make flashy promises to protect and compensate the original artists, which will most likely result in a few cents (or in the best case, a few dollars) received by individual artists for their images being used in AI model training. At the same time, they do support AI artists, providing them with the opportunity to sell their artworks.

But most importantly, they provide transparency in this new and underexplored art niche. When asking artists to explicitly label their images as Generative AI, they make sure that customers are fully aware that certain images on the platform are AI-generated images, and hence, can make an informed choice. Moreover, if in the worst-case scenario, the new legislation would recognize that AI-generated content violates the rights of original artists, whose works were used to train AI models, and would require banning AI-generated content from stock sites, Adobe Stock would be able to do this promptly – because everything is clearly labeled.

And what about stock sites that decided to ban AI-generated content? Well, they most likely still have quite a lot of this content on their website, but just not disclosed. There is simply no technology available at the moment that would allow recognizing AI-generated images with 100% confidence, except for some obvious examples with 6 fingers and other anatomical miracles 😀

So Can I Sell My AI Art?

Yes, you can! If you create high-quality images, you can start by putting them for sale on Adobe Stock. There are probably some other platforms that also support AI-generated art, but in this post, I could only cover the largest stock sites, and Adobe Stock looks like the best choice, at the moment.

Not, I am not paid for promoting Adobe. Unfortunately 🥲 

When it comes to AI tools that allow commercial use of generated images, you have a choice. At the very least, paid Midjourney users and DALL-E 2 users are allowed to use generated images for commercial purposes.

Here’s a great guide by Matt Wolfe on how to sell your AI images on Adobe Stock.

Next week, I’m going to look into major art marketplaces and their approach to AI-generated art. Stay tuned. And happy prompting! 😉 

🗞 News and Top Reads

• OpenAI released GPT-4 for ChatGPT Plus subscribers. This new, large multimodal model accepts images and texts as input and emits text output.

• The U.S. Copyright Office concluded that copyright does not apply to images generated by Midjourney. Their motivation is that users cannot control Midjourney’s output, in contrast to cameras or software like Adobe Photoshop whose output is subject to copyright. So right now, in the US, nobody owns the output of generative AI models.

• The Mauritshuis Museum in the Netherlands selected an AI copy of Vameer's Girl with a Pearl painting while it is on loan in another museum.

🏪 AI Art Tutorial of the Week

Hyperrealistic Midjourney Images - Complete Prompt Guide

In this video tutorial, Olivio Sarikas shares some great tips and tricks on how to achieve hyperrealism with Midjourney. You’ll see how to create hyperrealistic images in different styles across a variety of domains.

🎨 Artist of the Week: Refik Anadol

Refik Anadol (b. 1985, Istanbul, Turkey) is an internationally renowned media artist, director, and pioneer in the aesthetics of machine intelligence. He currently resides in Los Angeles, California, where he owns and operates Refik Anadol Studio and RAS LAB, the Studio’s research practice centered around discovering and developing trailblazing approaches to data narratives.

Residing at the crossroads of art, science, and technology, Anadol’s site-specific three-dimensional data sculptures and paintings, live audio/visual performances, and immersive installations take varied virtual and physical forms.

🖼 AI-Assisted Artwork of the Week

The woman in the painting is an embodiment of strength and courage. Her natural beauty is emphasized by the vibrant colors and her self-confidence is unmistakable. The woman looks self-sufficient and loves to adorn herself with flowers that she cleverly weaves into her hair. Her laconic yet proud stance hints at her position as a stewardess or lieutenant on a spaceship, giving her a mysterious allure.

Disclosures: The painting is based on sketches created with the help of Midjourney, and the text description was generated using Notion AI.

🤓 How to Get Started with AI Art?

1. DALL-E: Creating Images from Text – introduction to text-to-image generation.

2. The DALL-E 2 Prompt Book – a guidebook by OpenAI that explains how to effectively right prompts to generate images across different domains (e.g., photography, illustration, art history, 3D artwork).

3. Best Midjourney Prompts – a guide that covers the basics of Midjourney prompts (e.g., which keywords to use to create abstract art, surreal art, minimalism, etc) as well as some more advanced options (e.g., keywords related to camera lenses and filters, imitating certain artists and photographers without using their names). Finally, they provide a list of 600+ creative text prompts for image generation.

4. Stable Diffusion Prompt Book – a prompt book prepared by OpenArt. The book discusses ideal prompt format, using modifiers to change the style, format, or perspective of the image, applying ”magic words” to improve image quality, adding negative prompts, and adjusting Stable Diffusion parameters.

Share Kiki and Mozart

If you like this newsletter and know somebody who might also like it, feel free to share this newsletter. Let’s have more people learning about AI art!

If you have been forwarded this email and you like it, please subscribe below. And welcome to the world of AI art!