Adobe Unveils New ‘Firefly’ Suite of Generative AI Tools

Image: Adobe

Adobe today announced “Firely,” a suite of new generative AI tools to help customers create images and more. The company described Firefly as “a family of generative AI models for creative expression.”

While Adobe plans for Firefly to be an entire line of generative AI models, the first tools to be released to the public include an AI image generator like Midjourney or OpenAI’s DALL-E and a text effect generator.

“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,” said David Wadhwani, President of Digital Media Business at Adobe.

“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 said that models under the Firefly banner will be designed to serve customers with a wide array of skill sets and technical backgrounds and will cater to a variety of different use cases.

The first of Adobe’s Firefly models, an image and text effect generator, was trained on Adobe Stock images, openly licensed content, and public domain content. Adobe said that it is designed to generate content that’s safe for commercial use.

Users will even be able to train Adobe Firefly on their own datasets, making it possible for the models to generate content in the user’s personal style or brand language.

Adobe went on to note that future Firefly models will leverage a variety of assets, technology, and training data from the company and others. “As other models are implemented, Adobe will continue to prioritize countering potential harmful bias,” the company said.

Adobe is also implementing a new “Do Not Train” system to help mitigate the risk of abuse and intellectual property theft that AI image generators have been susceptible to since they entered the picture. The solution lets artists embed instructions in an image’s metadata that could prevent it from being included in datasets for training machine learning models.

“With industry adoption, this will help prevent web crawlers from using works with ‘Do Not Train’ credentials as part of a dataset,” Adobe explained in a separate blog post about responsible innovation in the AI space.

“Together, along with exploratory efforts to compensate creators for their contributions, we can build generative AI that both empowers creators and enhances their experiences.”

Firefly’s image and text effect generation tools are available as part of a public beta starting today. Both tools will be available through a website at launch, but Adobe eventually plans to integrate Firefly models directly into its creative apps, including Photoshop, Illustrator, and Premiere.

Anyone can sign up for access to the Firefly beta today — you don’t even need a Creative Cloud subscription. That said, Adobe will be limiting how many people it allows into the beta.

Microsoft earlier today launched an AI image generator of its own, dubbed Bing Image Creator, that’s powered by DALL-E. Google, meanwhile, made its AI chatbot and ChatGPT rival available to select users in the U.S. and U.K. today.