Databricks and Shutterstock unveil ImageAI to tackle copyright concerns in AI image generation

Databricks and Shutterstock have announced the launch of ImageAI, a new text-to-image diffusion model designed to eliminate the risk of copyright violations commonly associated with AI-generated images.

Unlike most image generators that rely on web-scraped content, ImageAI is exclusively trained on Shutterstock’s extensive image library.

Naveen Rao, Databricks’ head of generative AI, likens ImageAI’s approach to Apple’s resolution of the MP3 piracy problem in the early 2000s. “Apple solved this by figuring out a way to pay creators and providing a platform that enables people to buy music seamlessly,” Rao explains. He believes ImageAI will offer a similar solution by compensating content creators and providing businesses with a reliable source of AI-generated images without legal uncertainties.

The new model is hosted on Databricks servers, allowing enterprises to integrate the image generator into their applications via an API. This approach aligns with Databricks’ strategy of developing and hosting AI models within a secure data cloud environment. According to Shutterstock Chief Enterprise Officer Aimee Egan, both companies will sell access to ImageAI using their respective pricing models, with Shutterstock receiving payment for every generated image.

Watch the demo below to learn more about Shutterstock’s listing, and how to access and query it.

At launch, ImageAI does not support fine-tuning with customers’ brand imagery and style guides, though Egan suggests this feature will likely be added soon. This capability is highly anticipated by businesses that use AI-generated images for various marketing purposes, from print and social media to display and other media formats.

Rao underscores the importance of legal compliance in AI image generation. “I don’t want to get into legal waters by producing content that has Mickey Mouse in it by accident; I want to make sure everything that went into that is absolutely kosher in terms of permission,” he says. This need for upfront compliance is a significant draw for companies wary of potential copyright litigation.

ImageAI’s training on Shutterstock’s ethically sourced content marks a significant departure from the common practice of scraping images from the internet, a method that has led to numerous legal challenges.

By partnering with Shutterstock, Databricks aims to provide a model that respects intellectual property rights and compensates creators. This collaboration has already proven lucrative for Shutterstock, which earned $104 million in 2023 from licensing its image and video catalogs to various AI model developers, including OpenAI and Meta.

Shutterstock has been partnering with several AI companies in the last months. They announced a partnership with Reka.ai a week earlier and a new text-to-speech tool just days before that. Shutterstock also partnered with Nvidia for 3D models.