AI-driven image generators are currently facing a wave of copyright lawsuits, primarily due to their training on substantial datasets sourced from public websites. Many AI firms assert that the fair use doctrine protects their data scraping methods and training processes, though this viewpoint is contested by numerous copyright owners.
In response to these challenges, some startups and companies focused on developing image generation technology have opted for a different strategy: they are training their generators solely on licensed materials. One such company is Bria, which was established in 2023 by entrepreneurs Yair Adato and Assa Eldar, and has bases in New York and Tel Aviv.
Bria collaborates with around 20 partners, including Getty Images, to procure images that are utilized for training their image-generation models, all while adhering to content guidelines. According to Bria’s CEO, Adato, the platform compensates image creators in a “programmatic” way based on their “overall influence.”
“Our foundation models encompass one billion images and millions of videos,” Adato explained to TechCrunch. “We have reduced biases that might arise in AI-generated imagery by carefully training our models with data that represents diverse global perspectives. Our models routinely generate visuals that embody diversity, making them suitable for a range of creative uses.”
Bria provides plug-ins for popular image editing and design software like Photoshop and Figma, alongside a tuning API that enables clients to adjust their models for specific uses. Customers can operate Bria’s models on the company’s infrastructure or utilize external cloud services, with all data and outputs remaining the property of the users, as stated by Adato.
“Enterprise clients have the option to purchase access to our source code and models,” Adato remarked. “We offer over 30 specialized APIs for visual creation and modification, which clients can access via subscription and usage-based pricing. Businesses can invest in refining our generative AI models with their brand materials, crafting bespoke engines that preserve their visual identity.”
Bria has ambitious objectives. Adato mentioned to TechCrunch that the company, with a workforce of 40, aims to cultivate an “IP ecosystem” where enterprises can obtain licensed images from media giants for commercial projects, ensuring compliance throughout the process.
Additionally, Bria intends to broaden its platform and models to accommodate different media formats, such as music, video, and text, as well as applications that operate directly on devices.
“Bria continues to flourish amidst wider technological sector challenges,” Adato stated. “While the industry contends with issues stemming from the maturation of leading tech firms, economic pressures squeezing budgets, and an influx of simple AI applications, these factors actually bolster Bria’s standing.”
Numerous businesses are emerging in the quest to develop ventures around licensed media generators, with notable names like Adobe, Spawning AI, and Shutterstock entering the field. However, Bria has successfully carved out a niche in this evolving market. Recently, the company announced securing $40 million in a Series B funding round led by Red Dot Capital, with participation from Maor Investment, Entrée Capital, GFT Ventures, Intel Capital, and In-Venture.
With this latest funding, Bria’s total investment has reached approximately $65 million, with plans to allocate a significant portion toward product enhancement, as Adato noted.
“Our rapid growth is evident with our 40 clients, showcasing impressive annual recurring revenue growth exceeding 400% last year,” Adato mentioned. “We’re also expanding our workforce, bringing in more talent in key areas such as generative AI research in music and video, global sales and marketing leadership, IP and copyright legislation, and generative AI consultation. We anticipate doubling our team size by year’s end.”
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
