Art Marketplaces Respond to Generative AI
Image generated with the help of OpenAI DALL-E

Art Marketplaces Respond to Generative AI

Welcome to the Reality Gap newsletter, which focuses on synthetic data, generative AI, and data-centric computer vision. If you'd like to be notified about the next edition, click "Subscribe" at the top of this page.


Generative AI is causing a stir in online art communities.

As online art marketplaces offering photo and video content became overtaken by AI-generated images, artists are starting to turn against technology. AI art is seeing a tremendous amount of backlash.

Generative AI has been around for a while, but it gained popularity this summer after AI models that enable easy image creation using a text prompt, such as OpenAI DALL-E, Stable Diffusion, and Midjourney, became available to the public.

Almost overnight, many artists have started using generative AI technology to improve their creative workflows, while others are pushing back against it.

Some major online art marketplaces have banned AI-generated images

  1. Getty Images CEO Craig Peters announced a ban on the upload and sale of AI-generated illustrations on the site. Peters cited concerns about the legality of such content and a desire to protect the site's customers from potential legal risks.
  2. New Grounds, a website hosting user-generated content since 1995, also recently banned AI-generated art, stating a desire to keep the focus on art made by people.
  3. PurplePort, another major photography database company, also announced new rules regarding the removal of images created from text prompts on AI programs. The goal of the ban is to maintain a "safe, honest, and vibrant community" for creatives.

Enforcing the ban on AI art may be difficult, as there is no foolproof way to tell if an image was created by AI or a human. Websites will have to rely on users to identify and report AI-generated images, and these reports could be highly subjective and biased.

Additionally, there is a problem with defining AI art clearly. Many major image editing tools like Adobe Photoshop, Corel PaintShop Pro, and Canva incorporate generative AI capabilities into standard creative workflows. Some images may be fully produced by generative image models, while others may only be slightly aided by AI tools.

It can be challenging to differentiate between fully AI-generated art and art that simply incorporates AI assistance. Some may argue that a piece is only truly AI-generated if it is entirely created by the intelligent algorithm, while others may believe that any use of AI technology in the creative process qualifies as AI art.

Other art marketplaces take a less aggressive approach to generative art

Adobe has announced that it will now accept AI-generated content on Adobe Stock. Contributors will have to label AI-made content and obtain permission for any reference images or text used in its creation. Adobe Stock creators also need releases for any material depicting recognizable people. "We believe that generative AI tools can help our contributor community create amazing content," the company said.

Shutterstock, one of the largest stock image and video websites, announced it will ban any AI-generated art generated outside of its website but will soon integrate AI-generative capabilities as a part of its own toolkit.

No alt text provided for this image
A screenshot from Shutterstock.com

With the help of OpenAI, Shutterstock will train their own exclusive version of an image generative model on the corpus of images submitted by their members. This will make it a competitive advantage against other AI art providers.

“I think there are two choices in this world,” says Shutterstock CEO Paul Hennessy. “Be the blacksmiths that are saying, ‘Cars are going to put us out of the horse shoe making business,’ or be the technical leaders that bring people, maybe kicking and screaming, into the new world.”

In addition to offering AI-generated art, Shutterstock is also launching its Contributor Fund. This will be used to pay artists, photographers, and designers when their content is sold by the company to firms like OpenAI in order to develop generative AI models.

No alt text provided for this image
A screenshot from Shutterstock.com

It remains to be seen how the debate surrounding AI art will continue to evolve, but it is clear that the technology will continue to challenge conventional creative process

The arrival of generative AI has disrupted traditional art communities and business models. While some companies have banned AI-generated images, others have embraced the technology and found ways to push technology forward.

In the end, I believe that Shutterstock is pursuing a strategy that will be more beneficial in the long run. The genie is out of the bottle and it will be difficult to halt the progress of generative AI as it will present significant value to the public.

We will be monitoring this unfolding story and sharing updates in this newsletter.

And now, let's dive into the news headlines of recent days!


The Generative AI Revolution in Games

A great read by Andreessen Horowitz , a major VC firm, where they reflect on the potential impact of generative AI in the game development industry.

NVIDIA’s New AI: Generating 3D Models!

Two Minute Papers review 英伟达 's GET3D white paper "A Generative Model of High-Quality 3D Textured Shapes Learned from Images".

Datagen Introduces 50K Synthetic Identities Dataset

Datagen , a synthetic data generation platform/service provider, announced their new 50,000 synthetic identities dataset that is now accessible through the company's SDK and API.

Epic Games Releases RealityScan iOS App for 3D Scanning

Today Epic Games and Capturing Reality announce that the RealityScan app is now available to download for free on iOS. RealityScan is a 3D scanning app for mobile devices that easily turns photos into high-fidelity 3D models, with no experience necessary to get started.

Comparing Synthetic and Real Image Datasets Using UMAP

Rendered.ai has built an analysis method for assessing the relevant feature similarity of synthetic image data to real data of the target domain. The method uses UMAP dimensionality reduction to view image embedding features as a three-dimensional plot.

Synthetic Media Companies Still Hiring


And that's a wrap!

Here are a few more ways you can learn about synthetic data and generative AI:

See you next week!

Andrey

This is answer from OpenAI chatbot: "It is not possible to completely stop the production of AI-generated web content, as the technology behind it continues to advance and become more accessible. However, it is possible to properly manage the change and ensure that AI-generated content is used responsibly and ethically. This can involve implementing guidelines and standards for the use of AI in content creation, as well as providing education and training to those who are using the technology. It is also important to regularly assess and evaluate the use of AI in content creation, in order to identify and address any potential issues or challenges that may arise. Overall, while it may not be possible to stop the production of AI-generated content altogether, it is possible to manage the change and ensure that it is used in a way that benefits society as a whole." And now me :) From an economic point of view, the price of the output, i.e. digital content, will approach the price of the cost of producing the same with margins included. This means that sooner or later a long-term equilibrium will be reached that can no longer be stopped, but only slowed down..

Ryan Swanstrom

Product Content Creator

1 年

Clay Campbell maybe of interest to you

要查看或添加评论,请登录

Andrey Shtylenko的更多文章

社区洞察

其他会员也浏览了