Shutterstock pioneers “research license” model with Lightricks, lowering barriers to AI training data

Shutterstock is reimagining how AI companies access training data through a new “research license” approach, first by partnering with AI creative technology company Lightricks. The partnership announced today will enable Lightricks to use Shutterstock’s extensive HD and 4K video library to train its open source video generation model LTXV.

The new licensing model addresses a key challenge in AI development: the high cost of accessing high-quality training data. This allows companies to start with small research licenses for testing and experimentation before signing up for more expensive commercial licenses.

Making ethical AI development more accessible to startups

“Many companies and model trainers have chosen the path of fraudulent data scraping instead of making the necessary investments to achieve the quality and confidence levels needed to develop commercially viable models. ” said Daniel Mandel, Global Head of Data Licensing and AI at Shutterstock. said in an exclusive interview with VentureBeat. “However, we believe that financial investment should not be a barrier for those seeking to enter this field with an ethical approach.”

This two-step approach could change the way startups approach AI development. Craig Andrews, Global PR Manager at Lightricks, calls this “a great opportunity for small, agile developers looking to explore innovative applications of generative AI without the large upfront costs of traditional licensing. “It’s a turning point for people,” he said.

Legal protection and fair compensation in the AI ​​era

The timing is important as AI training data practices come under increasing legal scrutiny. Several major AI companies are facing lawsuits for allegedly misusing copyrighted material to train models. Shutterstock’s approach provides a legitimate alternative while ensuring content creators are compensated.

“We are setting the standard for ethical AI development while ensuring creators are compensated fairly for their work,” Andrews explains. “This approach not only fosters trust in the creative ecosystem, but also establishes a sustainable framework for responsible AI innovation.”

Revenue sharing: A win-win for creators and AI companies

Shutterstock has implemented a revenue-sharing model where contributors receive 20% of the revenue generated from data licensing deals. Contributors can also opt out of having their content used for AI training, but Mandel notes that only about 1% choose to do so.

Lightricks plans to use licensed video data to power LTXV, an open source video generation model released last month. The model has already received a lot of attention, with “thousands of downloads on GitHub and Hugging Face,” according to Andrews. One notable use case is real-time video generation for interactive e-commerce.

The partnership aims to solve technical challenges in AI video generation, particularly motion consistency in long videos. “One of the biggest technical hurdles in AI video generation is achieving consistent motion and structure across long video segments without sacrificing quality,” says Andrews. “Shutterstock’s high-quality video library provides a wide range of datasets to help address this challenge.”

For Shutterstock, this partnership represents a strategic shift in its business model. The company has already established partnerships with leading AI companies such as Nvidia, Meta, and OpenAI. Mandel emphasized that the research licensing model will allow smaller organizations and research institutes to access high-quality training data.

Setting a new industry standard for ethical AI development

The partnership also reflects the growing trend for transparency and ethical considerations in AI development. Lightricks has open sourced LTXV to foster collaboration and innovation. Meanwhile, Shutterstock’s licensing approach ensures that content creators are appropriately compensated.

“The key message here is that no matter the size or funding of your company, there is no longer an excuse to collect unlicensed content for training purposes,” Mandel concludes. “There are better ways to enter this evolving market.”

The partnership could set new standards for how AI companies access training data and could impact industry practices amid growing concerns about the sources of AI training data. The success of this model may determine whether other content providers follow Shutterstock’s lead and create more flexible and accessible licensing options for AI development.