This Tuesday, October 28 in Los Angeles, Adobe unveiled an in-depth review of its creative ecosystem. The company is further pushing the integration of artificial intelligence, no longer just as an accessory tool, but as an assistant more deeply integrated into its flagship software: Photoshop, Premiere Pro and Firefly. Here is a summary of the announcements.
The new era of “AI assistants”
If AI has been at the heart of Creative Suite’s flagship software for several years now, Adobe’s main announcement this year is what the company calls its “AI assistants”: conversational intelligence capable of understanding requests in natural language.
Specifically, a user can describe what they want to do in their own words (for example: “removes the background and makes the sky brighter“) and the assistant performs the corresponding actions, leaving the possibility of taking control manually at any time.
These assistants don’t just execute commands: they can also suggest creative ideas, give personalized advice, or even take care of repetitive tasks. Adobe will integrate these AI agents directly into its software such as Photoshop, Firefly or Express. The goal is to make creation more accessible and seamless, even for those who don’t master all the technical features.
Photoshop: even more artificial intelligence tools
Artificial intelligence continues to shape Adobe’s image editing software.
In terms of improvements, “generative fill”, the tool that allows an area of an image to be filled, eliminated or enlarged using AI, gains in speed, precision and can now collaborate with other artificial intelligence models such as those from OpenAI or Google.
In terms of news, Adobe officially launches “generative upscale”, a tool that instantly improves image resolution up to 4K. This feature allows you to multiply the resolution of an image (for example, transform a blurry or small photo into a sharp, detailed 4K image) while recreating missing details consistently.
The “harmonize” function will help to naturally insert an object or person into another image. Imagine that you have a photograph of a person taken indoors and you want to place it in an outdoor landscape. Normally, you would have to adjust the light, colors, shadows and perspective to make everything look natural. With “harmonize,” all of this is done automatically, thanks to AI.
Premiere Pro – Simplified Creative Editing and YouTube Partnership
On the video side, Premiere Pro gets a set of automation tools for creative editing, like smarter object tracking. But the feature that attracted the most attention at the Los Angeles Convention Center was “AI Object Masking”: a tool that automatically selects and isolates people and objects in a video to apply effects or correct colors without manual rotoscoping.
The main novelty in video this year is the partnership between Adobe and YouTube. This announcement follows the recent launch of Premiere Pro Mobile, a redesigned version for filming, editing and publishing from a smartphone. On the occasion of MAX 2025, Adobe announced the integration of a “Create for YouTube shorts” space directly into the mobile application. Creators will be able to edit, edit and publish their Shorts videos without leaving Premiere Mobile, with access to exclusive effects, transitions and templates designed for the vertical format and uses of YouTube Shorts, a platform that continues to grow in popularity.
Firefly: an AI toolbox for audiovisuals
Finally, Adobe introduced a new generation of Firefly, its all-in-one platform dedicated to AI-assisted creation. It now brings together the best artificial intelligence models on the market, its own and those of partners such as Google, OpenAI, ElevenLabs or Topaz Labsto offer a complete environment for visual, sound and video creation.
The platform integrates complete audiovisual production functions for the first time. Now you can automatically generate royalty-free music synchronized with a video or create realistic, multilingual voiceovers, with emotion and rhythm control.
Adobe emphasizes that its own AI models are “commercially safe”, meaning they can be used in productions without annoying consequences, as they are trained with authorized data and respect copyright. A key argument in the face of the numerous controversies over intellectual property in generative AI.
Source: BFM TV
