Adobe adds AI chatbots to Photoshop, Premiere AI masking, and Generative Audio in Firefly

Adobe has recently announced several new AI features across most of its Creative Cloud products, with new AI chatbots in Adobe Photoshop and Adobe Express among the most notable additions, allowing users to automate repetitive editing tasks and get help through conversational interfaces. The new Photoshop AI Assistant, currently in private beta, is integrated into the sidebar and can manage background removal, color changes, object selection, masking, and color edits, in addition to offering creative suggestions and tutorials. Users can ask the assistant to perform certain tasks through the “agentic” mode, where the AI interacts with the program’s interface in real time, allowing them to switch back to manual controls at any moment.

Aside from the AI chatbot, Photoshop also receives new Generative Fill partner models, including Google Gemini 2.5 Flash, Black Forest Labs FLUX.1 Kontext, a Generative Upscale feature powered by Topaz Labs AI, and the new Firefly Image Model 5 beta, Adobe’s most advanced model yet, capable of generating realistic 4-megapixel images without upscaling, along with a new Harmonize tool for realistic lighting and color blending.

For Premiere Pro, Adobe is adding AI Object Mask for automatic subject isolation, along with new rectangle, ellipse, pen, and vector masking tools for faster tracking. The app will also integrate with YouTube Shorts through a new “Create for YouTube Shorts” hub launching soon in both Premiere and YouTube, offering exclusive templates and effects. Adobe Lightroom introduces AI Assisted Culling in beta, a tool that helps quickly select the best photos from large collections by filtering images based on focus, angle, and sharpness.

The company is previewing Project Moonlight, an experimental cross-app AI assistant that connects Photoshop, Premiere, and Lightroom into a unified creative system that links to Creative Cloud libraries and social channels to learn a user’s style, generate tailored images, videos, and posts, and suggest growth strategies based on audience data. Meanwhile, Adobe Firefly adds new AI Generative Audio tools similar to Suno AI, including Generate Soundtrack and Generate Speech, both trained on licensed material for commercial use and powered by Adobe’s Firefly Speech Model and ElevenLabs. The company also announced it is developing a web-based Firefly video editor that combines all these tools and will enter private beta next month.