Adobe Launches AI Assistant in Photoshop and Updates Firefly

Adobe has announced the start of beta testing for its AI assistant in Photoshop, which is now available in the web version and mobile apps. Simultaneously, the developer is expanding the capabilities of neural networks in its content generation and editing tool, Firefly.

The voice assistant for Photoshop was first announced back in October at the MAX event. Now the feature is available to users: it can be used to remove objects or people, change color schemes, and adjust lighting based on text prompts. Simply describe the task in natural language—for example, ask to add a soft glow, crop the image to a specific format, enhance shadows, or completely transform the background.

Adobe clarifies that paid Photoshop subscribers will be able to use the AI feature without limitations until April 9, while users of the free version will receive 20 generations to start.

Another new feature, available in the public beta, is called AI markup. It allows users to draw on the image with a marker, and the neural network automatically transforms the marked areas. For example, you can sketch a flower or indicate an object to be removed from the photo.

Furthermore, Adobe is expanding Firefly’s functionality. The tool has received generative fill (previously available in Photoshop), which allows you to replace or add objects while adapting the background. New features also include:

• Generative object removal,
• Image expansion using AI,
• Image upscaling,
• One-click background removal.

The company notes that since February, Firefly subscribers have been able to generate content without limits—this was done to stimulate activity. In addition to its own models, Adobe has integrated more than 25 third-party solutions for video and image generation, including Google Nano Banana 2, OpenAI Image Generation, Runway Gen-4.5, and Flux.2 Pro from Black Forest Labs.