After signing the OpenAI contract, Shutterstock is releasing a generative AI toolkit that generates images based on text prompts. • TechCrunch


When Shutterstock and OpenAI announced a partnership to help develop OpenAI’s Dall-E 2 artificial intelligence image generation platform, along with the Shutterstock library to train and feed the algorithm, The stock photography and media giant has also hinted that it will soon bring self-generated AI tools to users. Today the company has packaged that product. Customers of Shutterstock’s Creative Flow online design platform can now create images based on text prompts. OpenAI and Dal-E 2.

The essence of this seemingly brandless feature is that Shutterstock says “ready to license” right after the image is created.

This is because Getty Images, one of Shutterstock’s biggest competitors, is currently embroiled in a lawsuit against Stability AI (the maker of another generative AI service called Stable Diffusion) for using images to train AI without permission from Getty or the rights holders. It has an important meaning.

In other words, Shutterstock’s service not only embraces the ability to use AI rather than the skills of human photographers to build the images you want to discover, but it’s also setting the company against Getty in some ways. Embrace the brave new world of artificial intelligence.

Stability AI has received significant funding, but not as much as OpenAI, which closed a massive $10 billion round as of yesterday and expanded its partnership with Microsoft.

In addition to working with OpenAI, Shutterstock earlier this month announced an expansion deal with Meta, the parent company of Facebook, Instagram and WhatsApp. This agreement (similar to OpenAI) will use Shutterstock’s library of photos and other media (including video and music). Build AI datasets and train algorithms. As a result, we can expect more generative AI tools to be released.

What’s interesting is that we don’t know the financial terms of deals with OpenAI, Meta, or our other partner, LG, but these services have clear commercial endpoints. Shutterstock appears to be worth jumping into and getting involved with these new technologies, and seems to be trying to build a business around them rather than letting them be cannibalized by these tools.

The big question is whether what Shutterstock offers is clearly differentiated enough from other offerings of generative AI tools for image creation and has a unique selling point. Yes, licensing is one of the attractive aspects now, but in the long run, what will differentiate them from each other if everything is built on the same platform? With Image Libraries, the idea is that for paying customers (and photographers who upload images), you can simply have better choices, better prices, better discoveries, and an overall better experience. Will these parameters stay the same in the AI ​​world or will they be erased?

To be fair, Shutterstock is pitching itself as an “ethical” partner here, promising to pay artists whose images are used to provide these new services. Again, the question is whether these payments come close to the compensation those artists and photographers would receive for providing the images.

“Shutterstock has developed strategic partnerships with key industry players such as OpenAI, Meta, and LG AI Research over the past two years to facilitate their generative AI research efforts, and now we are uniquely providing our customers with responsibly produced generative AI capabilities. You can,” Shutterstock CEO Paul Hennessy said in a statement today. “Our easy-to-use, generative platform will transform the way people tell their stories. You don’t need access to a creative team to create great work, our tools are built on an ethical approach and a library of assets representing the diverse world we live in, and we recognize and reward artists who contribute to the development of these models. ”

Leave a Comment