Adobe has introduced three key AI-powered features in the latest beta version of Photoshop. Powered by Firefly’s generative AI, these new tools aim to make image editing faster and more efficient for both professional and amateur users. Among the highlights of the update are “Generative Upscale,” “Harmonize,” and an improved object removal tool.

Photoshop Introduces New AI Tools

Generative Upscale, introduced in the beta version, can improve the quality of low-resolution images up to 8 megapixels. Users are offered fixed scaling options of 2x, 3x, and 4x.

Developed specifically for content creators and visual designers working with old digital photos, this tool provides increased clarity, sharpness, and resolution in images. Adobe stated that this feature was developed in response to popular demand from the Photoshop user community.

Upscale is available as an option accessible from the Image menu. Currently, the tool has a width and height limit of 4096 pixels, and if these limits are exceeded, it suggests new size options. The tool is designed to operate with minimal user intervention, and the goal is for results to be directly usable on print or digital platforms.

Another new feature, the Harmonize tool, allows objects in an image to blend more realistically with the background. Photoshop automatically adjusts lighting, color, shadows, and detail during this process.

The Harmonize tool, available in beta on desktop and web platforms, is now available in early access for iOS users. The feature works thanks to the integration of the Adobe Firefly Image Model.

Adobe has also improved its existing object removal tool. The new version allows for more precise object removal from an image and more natural-looking content.

This improvement enables cleaner and more convincing results, even in complex areas of photos. The object removal feature is available on both the desktop and web versions.

The update also includes the ability to switch between Firefly Image 1 and Firefly Image 3 models. This change provides new options for users who want to work with different AI models, particularly in the Generative Fill and Generative Expand tools. Model selection is only actively available in the desktop version.