Adobe boosts Creative Cloud with next-gen AI features at MAX 2025

Adobe made a sharp turn in its artificial intelligence (AI) approach Tuesday, allowing creators to use competing AI models directly inside Photoshop and other Creative Cloud apps. The announcement came during the company’s MAX 2025 conference in Los Angeles, alongside many new, shiny AI features.

Photoshop’s Generative Fill feature now works with Google’s Gemini 2.5 Flash and Black Forest Labs’ FLUX.1 Kontext models alongside Adobe’s own Firefly. Users can pick which AI engine handles their creative requests, from adding mountains to backgrounds to generating entire scenes.

The company also released two new audio tools for video creators. Generate Soundtrack creates licensed instrumental tracks that automatically sync with uploaded footage. Generate Speech turns text into voiceovers in more than 20 languages, with controls for speed, pitch, and emotion.

“We wanna help users prompt music,” Alexandru Costin, Adobe’s generative AI head, told The Verge. “It’s a new muscle we need to develop, so in order to make that easier and more accessible, if you give us your clip, we will predict what type of music goes with that clip.”

The Generate Soundtrack tool uses a “Mad Libs” style interface where creators select from preset options like lo-fi, hip-hop, or classical. They can also describe the desired mood in a text prompt. Each request produces four audio clips, with a five-minute maximum length.

Both audio features entered public beta in the Firefly app on Tuesday. Adobe trained its audio models on purchased content from rights holders.

“We purchased music and voice from IP owners, that’s why we have the confidence to offer it as commercially safe,” Costin said.

This copyright protection gives Adobe an edge over competitors like Suno and Udio, which face lawsuits for training AI on copyrighted material.

Adobe is building a new web-based video editor inside Firefly. The tool combines a multi-track timeline with AI generation features for voiceovers, soundtracks, and titles. A private beta starts next month, with a waitlist now open for early access.

The company also rolled out Firefly Image Model 5, which generates photos at native 4-megapixel resolution. The model supports “layered image editing,” allowing users to move, resize, and replace elements in uploaded images via natural language commands.

AI assistants are coming to multiple Adobe apps. Adobe Express now has a public beta chatbot that creates or edits designs from text prompts. Photoshop’s web version will get a private beta assistant that handles repetitive tasks like removing backgrounds from batches of photos.

Adobe teased early talks with OpenAI about integrating creative features directly into ChatGPT. No timeline was given for this potential partnership.