Thursday, February 22, 2024

This startup aims to train Art-Generating AI exclusively on Licensed Images

Share

The rise of generative AI, mainly text-to-image AI, has led to aa rise in both investment and legal challenges. Two companies, Midjourney and Stability AI, are at present facing lawsuits accusing them of infringing on the copyrights of artists by training their AI tools on scraped images. Moreover, Getty Images has taken Stability AI to court for allegedly using its images without permission to train its AI art generator. These legal challenges showcase the potential issues with generative AI that includes the risk of reproducing biased or unethical content. However, a new startup called Bria claims to address these concerns by training their image and video-generating AI in an ethical manner to minimize risk.

“Our goal is to empower both developers and creators while ensuring that our platform is legally and ethically sound,” Yair Adato, the co-founder of Bria stated.

“We combined the best of visual generative AI technology and responsible AI practices to create a sustainable model that prioritizes these considerations.”

Bria, the visual generative AI platform, was co-founded in 2020 by Avi Adato, who developed a passion for computer vision during his Ph.D. studies. Bria focuses to provide a solution to the complex and manual process of creating visuals by digitizing and automating the entire process through AI. Not similar to other text-to-image AI art tools that are available on the market, Bria mainly focuses on enterprise clients and was built with ethical considerations in mind from the beginning. The company’s other co-founder, Assa Eldar, joined in 2022, and Bria leverages recent advancements in AI and open-source models to provide an innovative solution to enterprises.

“I realized that there’s a real business use case for this,” Adato said. “The process of creating visuals is complex, manual and often requires specialized skills. Bria was created to address this challenge — providing a visual generative AI platform tailored to enterprises that digitizes and automates this entire process.”

Bria’s web app and Nvidia’s Picasso cloud AI service offers businesses with an image-generating AI platform to generate visuals for their social media posts, ads, and e-commerce listings. Customers can now use the platform to generate, modify, or upload visuals and use the “brand guardian” feature to confirm that their visuals comply with their brand guidelines. The AI is trained on authorized datasets, which Bria licenses from individual photographers and artists, media companies, and stock image repositories. Bria shares its revenue with these data owners, and the model rewards them based on their contributions’ impact, which allows artists to set prices for each AI-training-run basis. While other companies, such as Shutterstock, Adobe, and OpenAI, are exploring revenue-sharing models for generative AI, Bria’s approach is more extensive, according to co-founder Avi Adato.

Adato explains: “Every time an image is generated using Bria’s generative platform, we trace back the visuals in the training set that contributed the most to the [generated art], and we use our technology to allocate revenue among the creators. This approach allows us to have multiple licensed sources in our training set, including artists, and avoid any issues related to copyright infringement.”

Bria also clearly denotes all generated images on its platform with a watermark and provides free access or so it claims, at least to nonprofits and academics who “work to democratize creativity, prevent deepfakes or promote diversity.”

Bria plans to inaugurate an open source generative AI art model in the coming months, which will be including a built-in attribution mechanism. Whereas other attempts, such as Have I Been Trained? and Stable Attribution, aim to identify which art pieces contributed to a particular AI-generated visual, Bria’s model will allow generative platforms to establish revenue sharing arrangements with creators. However, the launch of the generative AI industry makes it even more challenging to evaluate Bria’s technology fully. It stays fuzzy how Bria discovers back visuals in training sets and uses this data to assign revenue. Additionally, there might be concerns from creators about unfair payments or bugs in the system leading to some creators being overpaid. Ultimately, only time will tell how Bria addresses these concerns.

Adato exudes the confidence you’d expect from a founder despite the unknowns, arguing Bria’s platform ensures each contributor to the AI training datasets gets their fair share based on usage and “real impact.”

“We believe that the most effective way to solve [the challenges around generative AI] is at the training set level, by using a high-quality, enterprise-grade, balanced and safe training set,” Adato said. “When it comes to adopting generative AI, companies need to consider the ethical and legal implications to ensure that the technology is used in a responsible and safe manner. However, by working with Bria, companies can rest assured that these concerns are taken care of.”

While Bria’s open source generative AI art model with a built-in attribution mechanism provides new revenue-sharing opportunities to creators, there are still many open questions regarding its implementation. For example, it’s unclear how Bria is tracing back visuals in the training sets and how they will resolve disputes from creators who feel underpaid or overpaid. In addition, content creators might need to take few extra steps to abolish their art from Bria’s platform, which shall be a burden. Although, Bria’s co-founder, Adato, declares that creators can opt-out of the platform and doesn’t see it as a significant issue.

“We’ve made it a priority to focus on safe and quality enterprise data collections in the construction of our training sets to avoid biased or toxic data and copyright infringement,” he said. “Overall, our commitment to ethical and responsible training of AI models sets us apart from our competitors.”

In spite of the competitive landscape and ethical concerns surrounding generative AI, Bria has secured $10 million in funding from bigshot investors including Entrée Capital, IN Venture, Getty Images, and a group of Israeli angel investors. This investment suggests that there is trust in Bria’s platform that focuses prominently on the enterprise and was built from the start with ethical considerations in mind. Though Bria faces severe competition from established players such as OpenAI, Midjourney, and Stability AI, as well as newer entrants like Jasper, the funding will help the startup keep up to develop its platform and transform itself in the market.

Adato said that Bria is currently serving “a range” of clients, including marketing agencies, visual stock repositories and tech and marketing firms. “We’re committed to continuing to grow our customer base and provide them with innovative solutions for their visual communication needs,” he added.

The promising success of Bria raises questions about whether it will inspire a new wave of generative AI startups with narrower scopes than established players, making them less unprotected to legal challenges. As funding for generative AI cools off due to upgraded levels of competition and liability concerns, it’s possible that more “narrow” generative AI startups could emerge and gain traction in the market.

Read more

Recommended For You