Adobe’s Firefly, a family of generative AI models, is being integrated with Google’s Bard, an experimental AI service. The announcement was made at I/O 2023.
Firefly allows users to produce content variations built on Adobe Stock images that have been deemed safe for commercial use and to change the image components while generating variations for email, mobile, web and other channels.
Firefly will become the premiere generative AI partner for Bard in the coming months, powering text-to-image capabilities, Adobe says.
In the first month since Firefly’s beta launch, Adobe’s creators used it to generate more than 70 million images with text-to-image, vector re-coloring and text effect capabilities.
Adobe has also expanded its Content Authenticity initiative.
This response to the Firefly beta “demonstrates the power and potential of generative AI to inspire more people to create and the strong demand for a creator-centric, commercially viable approach,” says Ely Greenfield, CTO, digital media at Adobe. “We’re empowering millions more people to use Firefly for creative inspiration and design and share standout content with Adobe Express through this integration with Bard by Google.”
Greenfield adds, "Generative AI has captured the world's attention and changed how we think about collaboration and productivity," says Sissie Hsiao, vice president and general manager of assistant and Bard at Google. The partnership with Adobe will provide users with “the power to bring their creative ideas to life, quickly and easily – directly in Bard," Greenfield says.
Adobe is also expanding its Content Authenticity Initiative (CAI). It will the CAI’s open-source Content Credentials technology to bring transparency to images generated through this integration.
“At this critical moment in history, as generative AI becomes more powerful and prevalent than ever, people need a way to tell what’s behind the content they’re consuming,” says Dana Rao, general counsel and chief trust officer at Adobe. “Content Credentials will enable creators to tell their stories authentically, while providing easy-to-use tools to verify how a piece of content was created and modified.”
The CAI now has 1,000 members, Adobe says.
One cautionary note has come from Yair Adato, co-founder and CEO of Bria, a generative AI visuals company.
Adato calls for an ethical attribution model that compensates creators appropriately while allowing the use of a substantial number of images to achieve the desired output quality.
Adato also states that the partnership shows that it is challenging for a single company to acquire enough licensed data for effective machine training.