The live Adobe Firefly (Beta) allows users to experiment with a text-to-image generator to create unique images from detailed text descriptions as well as effects that would apply styles or textures to text with a prompt.
Soon, users will also be able to recolor vectors to create unique variations of artwork from detailed text descriptions.
Other features currently "in exploration" include the following:
- Image Inpainting, which uses a brush to add, remove, or replace objects in an image;
- Personalized Results, which generates images based on your own object or style;
- Text-To-Vector, which generates editable vectors;
- Extend Image, which changes the aspect ratio of your image with a single click;
- 3D-To-Image, which generates images from the interactive positioning of 3D elements;
- Text-To-Pattern, which generates seamless tiling patterns;
- Text-To-Brush, which uses brushes for Photoshop and Frescro;
- Sketch-To-Image, which turns simple drawings into full-color images; and
- Text-To-Template, which generates editable templates.
Adobe Firefly is designed to be embeddable into existing workflows across Adobe Creative Cloud, Document Cloud, Experience Cloud, and Adobe Express.
Adobe also plans to make Firely available through Application Programming Interfaces or APIs on various platforms to allow customers to integrate into custom workflows and automation.
“Generative AI is the next evolution of AI-driven creativity and productivity, transforming the conversation between creator and computer into something more natural, intuitive, and powerful. With Firefly, Adobe will bring generative AI-powered ‘creative ingredients’ directly into customers’ workflows, increasing productivity and creative expression for all creators from high-end creative professionals to the long tail of the creator economy," Adobe digital media business president David Wadhwani said.
To ensure the responsible creation of AI-generated images, Adobe Firefly will automatically attach a tag in the embedded Content Credentials, easily distinguishing AI-generated art and content from original, manually created work.
The existing generative AI model is trained on a dataset of Adobe Stock, openly licensed work, and public domain content where the copyright has expired.
Adobe clarified that copies of customer content will not be included in Firefly models, as the company does not train on any Creative Cloud subscribers' personal content.
"We’re just scratching the surface of generative AI and every day, the technology is improving. As it continues to evolve, generative AI will bring new challenges and it’s imperative that industry, government, and community work together to solve them," Adobe wrote in a post, where it outlined the generative AI approach to ethical principles of accountability, responsibility, and transparency.
"By sharing best practices and adhering to standards to develop generative AI responsibly, we can unlock the unlimited possibilities it holds and build a more trustworthy digital space," it added.