Adobe’s generative AI can produce astonishing photos with users’ simple instructions,avoiding even copyright issues altogether.

Adobe's new Photoshop feature, Generative Fill, harnesses generative AI capabilities based on an artificial Intelligence model called Firefly.

Firefly acts as a creative co-pilot for users by learning from images collected by Adobe Stock and copyrighted photos.

Although Adobe’s long-loved Photoshop is a popular picture editing tool, its advanced editing features are difficult to handle, and results often show some signs of unnatural composition here and there.

Generative Fill, however, changes the aspect ratio of photos and fills the space with seamless naturalness.

With this feature, users can select desired areas using polygonal lasso tools and then the generative AI will bring up three different outcomes based on the users’ text instructions.

Transforming an original photo into a spontaneous-looking image is typically a very time-consuming and pain-taking process.

With the help of the Generative Fill, photo editing can become a simple task of choosing options and typing a few words

For example, if a user wants to increase the size of an image in a photo, they can use the Generative Expand feature to let AI expand the image according to their commands.

This AI feature outperforms traditional fill functions,providing users with creative and effective tools.

It is noteworthy that if users utilize Generative Expand on a blank screen without entering any commands, the Artificial Intelligence will create a natural-looking photo.

Generative Fill has garnered praise from many designers, editing-inevitable YouTubers, and reviewers who tested and reviewed the feature.

In addition, to prevent abusing the generative AI, creating content that may cause social controversies such as terrorism, racism, and gender discrimination is blocked.

Adobe's grand presentation, Adobe MAX Sneaks 2023

At Adobe's grand presentation named Adobe MAX Sneaks 2023 held from October 10 to October 12 2023, Adobe unveiled many remarkable technologies and features rooted on generative AI.

Difan Liu, a research scientist at Adobe, presented Project Glyphe Ease.

The presenter showcased Adobe’s unreleased version of Adobe Illustrator generating fonts by drawing a few alphabets.

The generative AI generated a font with a four-letter word written on a notebook.

Another research scientist Yi Zhou introduced “Project Poseable” mounted with generative AI.

First, she automatically made a 3D model of a woman sitting on a chair by pressing a button. Then, she uploaded a picture of a boy with a Kung Fu pose.

The 3D model of a woman instantly took the exact same posture.

Also, Generative AI created a picture with the 3D model when the presenter typed prompts and hit the ‘Generate’ button.

With the help of Project Poseable, making 3D models with certain poses becomes a much easier task. This feature is intended for normal people who have never considered making 3D models on their own.

Project Dub Dub Dub is an AI project that translates one language into 70 different languages and 140 dialects.

During the presentation, a senior research scientist, Zeyu Jin, demonstrated translating his speech in English into one in Spanish in just 5 seconds.

He uploaded a video of a woman speaking in Japanese, then he changed the language into six languages including English and Korean.

This technology is expected to break language barriers and make it easier for ordinary people to produce or dub video in various languages.

Chaemoon Han

Grade 9

Laurel Springs School