Adobe introduces an AI-powered assistant for Photoshop

Adobe announced on Tuesday that its AI assistant for Photoshop is now rolling out in beta to users on the web and across the mobile apps. At the same time, the company is introducing additional AI-driven image-editing features inside Firefly, its platform for media creation and editing.

The creative software company first revealed the Photoshop AI assistant during its MAX event in October. The feature, which is now beginning to reach users, can help with tasks such as removing objects or people from photos, changing colours, or adjusting lighting through text prompts. Users can also rely on natural-language instructions to tell the AI assistant to add a soft glow, crop an image to a specific format, improve shadows, or change the background to create a different overall look.

Adobe said paid Photoshop users will be able to generate unlimited edits with the AI assistant through April 9, while free users will begin with 20 generations.

The company is also launching a new feature called AI markup in public beta. This tool lets users draw markers directly on the screen and then use the AI assistant to transform those marked areas. For example, someone could sketch a flower or highlight an object they want removedtoo change the background.

In addition, Adobe is bringing new image-editing capabilities to Firefly, its media-generation product. Firefly is gaining Generative Fill, a feature already available in Photoshop for several years, which can replace or insert objects and adjust the surrounding background accordingly.

Firefly is also receiving a generative remove tool to remove objects from images, a generative expand feature to enlarge images with AI, and generative upscale capabilities. On top of that, Adobe is adding a one-click option for removing image backgrounds.

The company said in February that it was allowing unlimited generations for Firefly subscribers to encourage greater use of the service. Over time, Adobe has also expanded Firefly by adding more than 25 third-party video and image-generation models, including Google’s Nano Banana 2, OpenAI’s Image Generation, Runway’s Gen-4.5, and Black Forest Labs’ Flux.2 Pro.