BRIA Launches Legally-Trained Visual AI Foundation Models

BRIA, a company specializing in visual generative AI, announced today the release of a suite of visual generative AI foundation models that have been trained exclusively with licensed content from stock photography agencies they've teamed up with.

This new, responsible generative AI technology seeks to be a solution for creators and artists to be properly compensated for using their work in training datasets and for AI-generated images to be commercially viable – two of the most pressing issues surrounding visual AI usage. 

Are you intrigued? Read on!

BRIA Responsible AI Generative Models: A Collaborative Development

The recent announcement of BRIA's newest, patent-pending generative AI models discloses that they were built with the collaboration of multiple partner companies.

The models have been developed and trained from scratch by BRIA using Amazon SageMaker –BRIA is part of the AWS Partner network–, which makes it scalable and sustainable for commercial use.

The training datasets consisted solely of authorized, high-quality visuals from three high-profile stock media agencies: Getty Images, Alamy, and Envato.

The entire suite of models will be available on NVIDIA Picasso for developers to take advantage of this platform's scalability potential.

BRIA's intention is to provide AI engineers and creators with foundation models for visual generative AI that are ethically built and accessible for them to develop unique applications and tools capable of producing commercially viable, AI-generated media.

According to Co-Founder and CEO of BRIA, Dr. Yair Adato, their new suite of AI models “helps address (the main) obstacles (in generative AI), delivering businesses commercially viable outputs, based on permissioned data and most importantly, accessibility to any developer or business.” 

Strategic Partnerships with BRIA for Responsible AI Development

As we just saw, BRIA joined efforts with several leading companies from different industries with shared goals in AI in order to develop this new offer.

The renowned stock photography agency Getty Images has for a while been fully committed to enforcing and protecting artists' intellectual property rights while fueling the development of visual AI tools, and their partnership with BRIA was announced some time back. As Peter Orlovsky, Senior VP of Strategic Development at Getty put it, they believe in these generative AI models' “tremendous potential to unlock creativity, allowing customers to unleash their imagination to generate visuals that are still commercially safe to use.”

As part of the announcement of this new technology, Alamy‘s Managing Director, Emily Shelley, highlighted that the BRIA models' ability to compensate artists for their source material is what sets it apart and that they feel that matches the agency's core principle of democratizing photography. Envato, on their part, has stated their company is all about empowering creators to thrive and that their involvement in a sustainable and inclusive ecosystem for creative innovation is framed as part of their efforts to bring more opportunities for artists to earn from their work.

These are the main but not the only companies BRIA teamed up with in its mission to create a responsible visual AI industry; its other collaborations include boutique stock photo agencies and smaller stock photo agencies such as Stockphotos.com.

These image providers are all collaborating in BRIA's new Fair Diffusion program, an opt-in for both large libraries and individual photographers and artists that uses reverse technology to identify whenever an artist's work has been used to generate a synthetic image, providing a way for them to be properly credited and compensated for said use.

Regarding tech partners, Amazon Web Services thinks their collaboration with BRIA through SageMaker is “paving the way for businesses to unlock the full commercial potential of Generative AI without compromising quality,” in the words of Udi Katz, Director of Startups for AWS.

Last but not least, NVIDIA's VP of Platform Marketing, Jason Paul, explained the value of their involvement with the Picasso service: “With GPU-optimized inferencing and simple API-as-a-service workflow, the NVIDIA Picasso Generative AI foundry for visual design enables BRIA AI to rapidly scale deployment for broad adoption,”

Establishing Commercially Sustainable and Responsible AI 

BRIA's entire offer and their newest suite of generative models, in particular, highlight the most pressing issues in the novel visual AI field: producing commercially viable AI-generated media and doing it respecting artists' copyrights and livelihoods.

Their solution seeks to provide a way to develop AI image generators that comply with copyright laws, properly compensate creators for their work, and generate AI imagery that is legally sound to use for commercial purposes, all in an easily scalable manner.

But their goal goes beyond that. In the words of BRIA CEO Dr. Adato, their foundation models represent “a movement – an open call to collectively shape a sustainable future for Generative AI.”

THE AUTHOR

Ivanna Attie

All About Ivanna

I am an experienced author with expertise in digital communication, stock media, design, and creative tools. I have closely followed and reported on AI developments in this field since its early days. I have gained valuable industry insight through my work with leading digital media professionals since 2014.

• NETWORKING