Adobe is the latest tech giant to jump on the bandwagon of generative artificial intelligence. The American company, whose products revolve around creativity, has announced Fireflya generator of images and effects from text that is presented as a complement to its suite of applications.
The announcement lays the foundation for a new strategy at Adobe. Although the software developer has been using AI for a long time, it has not been until now that they have embraced the generative possibilities that it offers. However, he seems to want no part of the copyright discussion that has arisen.
Adobe’s AI model
AI imaging models have typically been trained with enormous datasets containing images with Copyright. These practices have aroused the rejection of many artists, who do not agree that their works are used to give life to the aforementioned tools. Adobe, for its part, has taken a different path, training its own AI model.
With the help of NVIDIA, the American giant ensures that it has fed Firefly with content with an open license, content in the public domain whose copyrights have already expired and with content from Adobe Stock. Regarding the latter, he assures that he is working on a compensation model so that creators who upload their work to the image bank receive a benefit for contributing to Firefly.
However, not all content creators may be in favor of the aforementioned model. In this sense, from Adobe they indicate that they will respect the “no training” tags included in the works. This is an interesting option, but it forces artists and designers to protect their work individually instead of allowing tags to be added that authorize its use in AI training.
Regarding the operation of the tool, the trained model has been vitaminized with what they call “a style engine”. It’s an elementary piece in the Firefly experience that promises to help you generate photos, art, graphics, text effects, and more with the ability to control style, color, lighting tone, and composition. The idea is that this will be added to the Adobe Creative Cloud family, Document Cloud, Experience Cloud and Adobe Express in the future.
Since Firefly is currently in beta, its general rollout has not yet occurred. It is currently available to some users through a waiting list. Those interested can sign up at firefly.adobe.com. Once they receive access, as we can see in the images, they will be able to unleash their creativity from a browser application. It is time to wait to see its arrival in the applications of the company’s suite.
In any case, users have more and more AI-powered generative tools at our disposal. And its evolution is amazing. Just look at how Midjourney has improved over time. Version 5 of the tool is a true wonder and surprises with its realistic results. But we also have alternatives like DALL·E, Stable Diffusion and Bing Image Creator.
Images: Adobe (press kit)
In Xataka: Voice assistants were once the great promise of AI. Today they are nothing more than the ugly duckling
