Shutterstock’s generative AI way forward: 6-year training data deal with OpenAI

Enterprise customers can reach out to their account representative to inquire about securing exclusive rights for the image like you can do for traditional stock images on enterprise plans. AI-generated images represent new content that is created using AI technology trained on millions of real content assets, descriptions, and keywords. AI content generators require some human input like a description, prompt, or parameters. Plainsight provides the unique combination of AI strategy, a vision AI platform, and deep learning expertise to develop, implement, and oversee transformative computer vision solutions for enterprises.

It’s a smart decision to buy photos from a reputable source because you never know when someone will claim copyright infringement on an image you thought you purchased lawfully. To Hennessy there’s no conflict between building generative AI products with Shutterstock’s first party data and licensing it to others to grow its total contract value. Other companies with extensive media catalogs have waffled between taking money upfront and licensing content to digital platforms and keeping it exclusive. If you want to know more about the moral component, see our compilation of the ethical pros and cons of AI art generation. We also take a deep dive into whether AI-generated art should be considered real art.

Browse News Releases

Please notify us of any inappropriate or offensive content by clicking the Feedback button and providing details of your experience. This is just one more cost of doing business when it comes to generative AI. Large companies are known for their brands, manufacturing plants, complex supply chains, and, in good times, generating a lot of revenue. Few of them will see generative AI as core to their business and will likely forgo its use to avoid lawsuits. In an interview with The Verge, Getty Images CEO Craig Peters expressed concerns about the legal risks surrounding the technology. As we all know, at times during the production process, some silences can also get recorded.

However, he thinks there could still be other ways to use the tech down the road that haven’t been discovered yet. The photo stock company says that it indemnified AI images “for a leading cable network’s upcoming television series” and is now extending that protection to all enterprise customers. Although scraping or buying data to train AI art generators seems to be legal (covered by Fair Use), many experts worry about future challenges and complications. Getty Images, for example, has banned the sale of AI art on its platform because of fears that its inability to copyright the output of these systems will lead to licensing problems for customers. This isn’t the first time Shutterstock and OpenAI have worked together in this domain. From 2021 onwards, Shutterstock sold images and metadata to OpenAI to help create DALL-E (OpenAI’s Altman describes this data as “critical to the training of DALL-E”).

What Is Generative AI?

Image-to-text generators, such as DALL-E, have been trained on millions of images that were scraped from the internet. While the image generator companies don’t say as much, the data set will include millions of copyrighted images. The takeaway is that Shutterstock sees generative AI as a way to expand its total addressable market. Shutterstock will use generative AI tools to drive subscriptions and licenses and improve its products. What remains to be seen is whether granting OpenAI access to Shutterstock’s first party data will be a business risk. Additionally, Shutterstock plans to compensate the artists whose images are used during the generative AI process by creating a cash fund and paying them royalties.

Unlike other AI image generators out there – with Shutterstock you don’t pay to explore the options, you only pay for the content that meets your creative needs. Generative image models are a new and impressive technology, but they are not yet perfect and sometimes get confused when creating precise details. Sometimes this includes attempts at rendering text in a style influenced by the text that appeared in some of the images the model was trained on.

Copyright Office (USCO) later reversed its decision, the implications of this milestone unearthed an ethical quagmire for visual artists, photographers and the platforms they host images on. Shutterstock has partnered with OpenAI, the team behind popular generative solutions like DALL-E 2 and ChatGPT, to introduce a new solution for its paying customers. The largest online repository of stock imagery now allows users to create images from text prompts in any of the languages it supports. Generative art AI models are trained on scrapped internet data, in this case, images. While some analysts have argued these systems are using this data under fair use, others have argued AI search engines are violating copyrights.

Shutterstock Introduces its Own Generative AI Solution

Yakov Livshits
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.

By incorporating Picasso’s latest feature, artists can enhance, and light 3D scenes based on simple text or image prompts, all with AI models built using fully licensed, rights-reserved data. Shutterstock currently has AI-powered capabilities built in its platform after the acquisition of three companies, Pattern89, Datasine, and Shotzr. These tools work primarily in assisting users with the image selection process. Its catalog has over 200 million royalty-free stock images and other graphics, as well as video clips and royalty-free music for licensing. Shutterstock’s indemnity program is similar to the one announced last month by Adobe, which unveiled its own program alongside the release of Firefly, a generative-AI-powered image creation tool.

• In addition to covering MarineSitu’s history and the origins of this exciting partnership, the two experts discuss some of the promising future use cases for AI-equipped, submersible monitoring solutions.
• Now this unique partnership has enabled creators to quickly create and customize 3D scene backgrounds with the help of generative AI, thanks to cutting-edge tools from Shutterstock.
• Urban landscapes and neon colors bring any generated content to life in a bold futuristic cityscape.
• Through the widest breadth of managed services and a vision AI platform for centralized processes and standardized pipelines, Plainsight makes computer vision repeatable and accountable across all enterprise vision AI initiatives.
• Recently, the platform has begun to embrace content made with artificial intelligence, moving beyond human-made photos to experiment with images not taken with a traditional camera.

Getty Images described the collection as “a new category of digital art” and said that it would be offering them for sale under a royalty-free license. Following the recent release of AI-generated images by stock photo sites Getty Images and Shutterstock, both of the industry’s top companies have now officially banned and removed the submission of AI-generated visual art on their platforms. The decision was made due to concerns that these images could lead to legal challenges in the future. 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. 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.

Cory Gunther is a staff writer at Valnet, covering cutting-edge EVs, automotive, and more at How-To Geek and MakeUseOf. When he’s not reviewing the latest gadget, you can find him off-roading in his Toyota truck, fixing said truck, or doing something outdoors with friends and family. He’s covered CES, SEMA, Mobile World Congress, and CTIA and reviewed countless phones, TVs, speakers, EVs, e-bikes, and mobile Yakov Livshits accessories. Prior to Valnet, Cory worked for GottaBeMobile, and his writing also appeared on SlashGear, AndroidCentral, and Newsweek, and he’s written over 9,000 articles. A historical comparison would be Starz, which licensed its catalog to Netflix in 2008. Suffice it to say that this lineup of experts will bring valuable insight into the present and the future of generative AI for the creative industry.

“We’re at an inflection point in the use of generative AI technology,” says John Lapham, General Counsel at Shutterstock. Multiple recent lawsuits have been levied against the AI art generator, Stable Diffusion, and others for copyright infringement. And there’s not yet a clear legal precedent for how these cases will be handled.

Shutterstock is Granting Legal Protection to its Generative AI Users

In a strategic move signifying its commitment to artificial intelligence (AI)-powered solutions, Shutterstock recently announced the widening and deepening of its relationship with OpenAI. The company intends to supply the AI startup with an array of training datasets over an extended timespan of six years. However, not every stock image website is eager to rush onto the generative AI train. Right after the Shutterstock announcement, competitor Getty Images said it would be partnering with AI company BRIA to develop “responsible AI” tools centered around image editing, but not generation.

The conversation ends on a cliffhanger, so be sure to check back in for Part 2. Get going with our crush course for beginners and create your first project. With the help of an AI tool, you can automatically replace and remove Yakov Livshits the unwanted background from your clips. Using this tool, you can add virtual background to make your content more appealing. This way, you can leave a positive impression on your audience and engage them for a long time.

Firefly, Adobe said, works through training on both images owned by the company, and those in the public domain or other material not subject to copyright rules. Like Shutterstock, Adobe said that its indemnification for AI-generated images is meant to be as similar as possible to the one that covers the company’s other assets. Stock content provider and creative suite Shutterstock is the latest company in its field to offer customers a legal indemnity against suits related to AI-generated images created and licensed on its platform. Unlike some stock image galleries, Shutterstock has embraced the potential of generative AI. By partnering with OpenAI, Shutterstock has launched an image creator powered by OpenAI’s DALL-E 2.