Shutterstock’s generative AI way forward: 6-year training data deal with OpenAI
This follows widespread criticism from artists whose output has been scraped from the web without their consent to create these systems. Notably, Shutterstock is also banning the sale of AI-generated art on its site that is not made using its DALL-E integration. The partnership signifies a shift in the stock content industry, with Shutterstock embracing generative AI. It showcases the potential of generative AI in creating highly customizable stock images. The new tool utilizes a cutting-edge generative AI model that has been trained on Shutterstock’s vast library of content. This allows the system to generate new images, videos, and illustrations that are similar in style and composition to existing content, but completely unique.
Shutterstock’s image generator is able to produce unique images from a single word input or short simple phrases and features an intuitive style picker along with support for over 20 languages. Shutterstock, one of the internet’s biggest sources of stock photos and illustrations, is now offering its customers the option to generate their own AI images. In October, the company announced a partnership with OpenAI, the creator of the wildly popular and controversial DALL-E AI tool.
Adobe Stock Starts Paying Contributors for Using Their Work on Firefly
This is a big win for Shutterstock users and the creative industry as a whole, as it opens up a new realm of possibilities for visual content creation. Popular stock photo sites Shutterstock and Getty Images have announced partnerships with generative AI platforms that will allow users to create original custom photos with new tools using text-to-image models. Yakov Livshits The AI image generator is trained on hundreds of millions of ethically-sourced assets, ensuring customers can generate and license new assets with commercial safety while protecting and compensating the artists who contributed to the content that trained the models. Shutterstock’s tools are built on an ethical approach and a diverse library of assets.
All AI is trained on datasets, i.e. massive aggregations of material that teach it what to aim for. And for AI image generators, those training sets contain images made by humans—often human artists for whom their work is their livelihood. In addition to textual content, AI tools can be used to generate characters, images, and illustrations to support textual content.
Shutterstock is Granting Legal Protection to its Generative AI Users
At the close of the last quarter of the 2022 fiscal year, the first payment to contributors was set to be made in December. It is unclear how many contributors were paid in December and how much was distributed. Add a Pro Photo Daily Membership today to get complete website access to all articles and all discussions. If satisfied, sellers can directly upload this content to their product listings. DALL-E’s competitor Stable Diffusion, owned by Stability AI, was hit with two lawsuits last week. The first is a class-action lawsuit by a group of artists who are also suing Midjourney.
- The first is a class-action lawsuit by a group of artists who are also suing Midjourney.
- However, text-to-image models will map words to usual representations of objects and scenes and are trained on massive datasets obtained from sites such as Pinterest, Artstation and Flickr.
- However, Shutterstock has stated that their AI image generator was trained on datasets licensed by the company.
- That has continued to raise questions about copyright, such as who owns the copyright of AI-generated images and the sources of training images.
The company added an indemnification clause to its license boilerplate text, asserting it will take on any legal claim against a user over copyright or other legal misuse because it is confident that its text-to-image generator doesn’t violate any of those regulations. The announcement comes shortly after Adobe made a similar pronouncement about its Firefly synthetic media generator a couple of weeks ago. While stability AI’s training dataset LAION is publicly available, OpenAI has not revealed what DALL-E is trained on. However, Shutterstock has stated that their AI image generator was trained on datasets licensed by the company. Both Shutterstock and Getty Images have banned AI-generated images from being uploaded to their websites and Shutterstock still bans images generated by AI from its website, except those created using its own tools. A top supplier of stock images will compensate artists who contribute training data to its image-generation service.
Shutterstock continues generative AI push with legal protection for enterprise customers
Founder of the DevEducation project
A prolific businessman and investor, and the founder of several large companies in Israel, the USA and the UAE, Yakov’s corporation comprises over 2,000 employees all over the world. He graduated from the University of Oxford in the UK and Technion in Israel, before moving on to study complex systems science at NECSI in the USA. Yakov has a Masters in Software Development.
What remains to be seen is whether granting OpenAI access to Shutterstock’s first party data will be a business risk. The photo stock giant has rolled out OpenAI’s text-to-image generator integrated into its website after the partnership was announced last October. Many of these artists are tapping generative AI to bolster their complex workflows — and will be able to use the technology to quickly create and customize environment maps. This allows more time to work on hero 3D assets, which are the primary assets of a 3D scene that viewers will focus on.
It’s obvious that AI must be pulling its “inspiration” from the work of real, live people. But it’s difficult to pin down exactly when and where AI generators steal from visual artists. On the other hand, AI’s Yakov Livshits acts of plagiarism are much more apparent—though no more egregious—in AI-produced text. Clearly, if not approached carefully, artificial intelligence could pave the way for a theft crisis in creative fields.
In just a few seconds, you’ll receive four unique images in a wide range of styles. While you might occasionally find strange images being generated, remember that generative AI is learning all the time. Look at the funky hands as a growing pain that only crops up every once in a while.
Each query results in the creation of four images, and the user is then able to further filter the type of image the AI creates. Images are free to generate but users must sign up with Shutterstock in order to download and use the images, with costs per image ranging from 48 cents to $2.90 per image, depending on the credit package the customer has purchased. Shutterstock’s move to incorporate generative AI into its platform is a significant step forward for the company and for the industry as a whole. It is expected to have a major impact on the way visual content is created and used, and will likely be a catalyst for the development of new and innovative tools and technologies in the future.
It’s a question asked by many following the rise of text-to-image AI models in recent years. The answer from the industry’s incumbents, though, is “no” — not if we can start selling AI-generated content first. Shutterstock’s chief executive officer, Paul Hennessy, said in the announcement that the AI generative platform will transform the way people tell their stories — without requiring the user to be a design expert or have access to a creative team to create exceptional work.
Headquartered in New York City, Shutterstock has offices around the world and customers in more than 150 countries. Shutterstock, which has more than 615 million images enriched with metadata and 2.2 million global customers, will provide training data for OpenAI models. Shutterstock and OpenAI launched a partnership in January, but the expanded deal will give OpenAI a license to access Shutterstock training data for image, video and music libraries and metadata. Contributors to stock image galleries, including artists and photographers, have expressed concerns about generative AI startups profiting off their work without providing credit or compensation. To address this, Shutterstock maintains a “contributor fund” that pays artists for the role their work has played in training Shutterstock’s generative AI and ongoing royalties tied to licensing for newly generated assets.
“Our tools are built on an ethical approach and on a library of assets that represents the diverse world we live in, and we ensure that the artists whose works contributed to the development of these models are recognized and rewarded.” Generative AI has revolutionized the creative industry by offering highly customizable and on-demand content. This technology allows AI models to create unique stock images, videos, and music in real-time, posing both opportunities and challenges for stock content galleries like Shutterstock. While generative AI empowers users with endless creative possibilities, it has also raised concerns among artists and photographers who fear their work may be used without proper credit or compensation. More importantly for the future of content creators and AI companies, Shutterstock also launched a Contributor Fund in October that is designed to reimburse content creators when the company sells their work to train text-to-image AI models.