American company Adobe presented new generative artificial intelligence Firefly, which can create images from a text description. The AI is currently in beta testing and can only be accessed by invitation. Adobe’s Firefly is reminiscent of Midjourney and DALL-E generative AIs already on the market.
To simplify the text description, Firefly will have built-in options for lighting, art styles, and aspect ratio. It will be possible to impose effects on an already finished image. Another tool will allow you to create decorated inscriptions. According to the developers, AI training took place on the basis of images without copyright infringement, including from the Adobe Stock library. This will not allow you to generate random combinations from already known brands or other people, as other generative AIs do.
Adobe also plans to allow artists to train Firefly on their own images, which will bring new styling options to AI. Going forward, the company plans to include Firefly in its products, including Photoshop, Illustrator, and Premiere Pro. Photoshop will have a generated fill, Premiere Pro will be able to change the style or color of an image, and Illustrator will generate vector variations of hand sketches.
