Adobe expands AI-powered Firefly into audio, 3D content

2023. 10. 12. 16:33
글자크기 설정 파란원을 좌우로 움직이시면 글자크기가 변경 됩니다.

이 글자크기로 변경됩니다.

(예시) 가장 빠른 뉴스가 있고 다양한 정보, 쌍방향 소통이 숨쉬는 다음뉴스를 만나보세요. 다음뉴스는 국내외 주요이슈와 실시간 속보, 문화생활 및 다양한 분야의 뉴스를 입체적으로 전달하고 있습니다.

2023 Adobe MAX [Photo by Ko Min-suh]
Adobe Inc. plans to expand the use of generative artificial intelligence (AI) technology across its entire product lineup, the world’s leading content creation software company said on Wednesday.

Starting last month, Adobe initiated the commercialization of its creative generative AI model family known as Firefly.

Under its plan announced at the 2023 Adobe MAX event held in Los Angeles on Wednesday, Adobe plans to have Firefly to cover not only images but also audio, video, 3D content, and more, making it a comprehensive tool for content creation.

As the first step towards the expansion, Adobe unveiled three new models under the Firefly umbrella – Firefly Image 2 Model, Firefly Vector Model, and Firefly Design Model.

These models have integrated AI capabilities into Adobe’s content creation tools, such as Photoshop, Illustrator, and Adobe Express, and they are now available for users to experience.

The most highly anticipated model among these is the Firefly Image 2 Model, which goes beyond the capabilities of the original Firefly Image, released in beta in March.

It introduces a generative match feature that allows users to create content with custom styles.

According to Adobe, users can either choose images from a predefined list or upload reference images when generating new images using text prompts through the feature.

Since the beta release of the Firefly Image model in March, more than 3 billion images have been generated, demonstrating the strong market response to this technology.

Adobe plans to sequentially introduce new generative AI models in the Firefly family, with future models targeting audio, video, and 3D content, according to David Wadhwani, president of the digital media business at Adobe.

One noteworthy feature of the Firefly generative AI models is their foundation in learning from Adobe Stock images, open-source licensed content, and public domain content with expired copyrights.

This minimizes the risk of copyright conflicts, setting Adobe’s Firefly apart from other generative AI models that often face limitations regarding commercial usage due to copyright issues.

Copyright © 매일경제 & mk.co.kr. 무단 전재, 재배포 및 AI학습 이용 금지

이 기사에 대해 어떻게 생각하시나요?