JAKARTA - The BRIA startup, which is responsible for AI's image generation and is backed by stock photo provider Getty Images, raised $24 million in the Series A funding round, including from the leading advertising agency Publicis Groupe, as announced on Wednesday, February 21.

BRIA, based in Israel, which contracted more than a billion images from stock providers to its systems, stated in a press release that it would use the funds for global expansion and expand the ability to create videos from text.

This investment, although small compared to others in the hot AI space, made BRIA a preliminary test case in how licensing agreements could provide a way out in an ongoing dispute over the use of copyrighted content to train a generative AI model.

Music labels, artists, and other content owners have designated licensing as a more acceptable approach to generative AIs, after big AI companies built their systems using content taken from the internet for free.

BRIA says its approach is only through licensing to prevent the risk of trademarks and other harmful elements from appearing in the images it produces, attracting customers in the world of advertising, marketing, and the media including publishers.

"There is a way to do it that is responsible and commercial," BRIA CEO Yair Adato told Reuters in an interview.

Make it, which has supported BRIA since 2022, has become an important part of the dispute surrounding training data.

Last year, Getty sued leading AI image company, AI Stability, for using his unlicensed photos, which they accused of violating "chrand" intellectual property rights on "admirable scale."

They also partnered with chipmaker Nvidia last year to build image generation service trained only on images in their own library.

Adobe and Shutterstock also offer image makers AI, while the three photo companies license their visuals to Nvidia for their own services.

BRIA is 'feeding' its models with visuals of 18 stock providers, including Getty, Alamy, SuperStock, and Envato. They charge customers to access the platform and share revenue with photo companies.

Whenever BRIA produces images, they display original license images that contribute to the output and record how many come from each other.

Adato likened this approach to Spotify's model for the AI era, which replaced content owners for their role in enabling the technology.

Controlled training also protects against the technology used to create misleading visuals that can damage brands or influence elections, he said.

"The BRIA model, for example, was unable to produce images of US President Joe Biden because it had never been taught what his face looked like," said Adato.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)