vAIsual Leaps into AI Video Generation Training with Biometric Video Datasets

The pioneer in ethical AI training datasets vAIsual, has recently included high-resolution and legally clean video datasets on their Dataset Shop, making them available for developers to license and use to train their AI video generative models. 

As AI text-to-video generators are one of the upcoming and most awaited developments in AI-powered tools (with some basic apps already available), having this professional and ethical resource ready is certainly an interesting advantage for AI startups.

Biometric Video Datasets in the Dataset Shop

The new addition to the Dataset Shop (a marketplace for visual media designed specifically for AI training purposes, owned by vAIsual) is a collection of 400+ videos in 4K quality with no post-processing, featuring over 100 different people. 

The clips are all shot with a blue screen background and include both full-body and portrait takes, covering a wide range of body postures and facial expressions as well as diverse ethnicities, ages, genders, and body types. 

Equally relevant, all the models featured in the videos have signed biometric model releases, agreeing to have their likeness used to train AI models, and all the clips are shot in-house and owned by vAIsual themselves, granting copyright permissions to those who use them in AI training. 

This kind of dataset allows developers to train their generative models with diverse, rich, and authorized content, resulting in tools that create legally safe(r) AI-generated videos. 

You can check out the biometric video dataset here!

vAIsual’s Datasets More Valuable than Ever

vAIsual is a company that anticipated the need for legally clean datasets to train AI generative models way before bigger and more popular players did. Their upper hand is undoubtedly linked to the fact the team behind it has decades of combined experience in the stock media licensing field and thus is highly educated in image copyright and all other rights that need clearer for the commercial use of images. They're renowned professionals in their field – for example, they'll be participating in AI-focused panels at the upcoming CEPIC Conference 2023 alongside our resident expert Amos Struck.

Known as “the clean data guys” in the generative AI space, their Dataset Shop remains a unique offering in AI training datasets with all rights cleared and all permissions in place to grant its legal safety. They launched it in 2022 with a collection of in-house datasets (which at the time was the largest pool of biometric still images) and have since expanded through partnerships with stock media agencies. 

Their goal is to stop developers from scraping visual content off the web and training AI models with unauthorized, copyrighted content, giving them a professional, readily available, and more ethical alternative. And this offer becomes increasingly more relevant as time goes on and the legal frame for AI-generated media gets better defined.

Now we know that AI images might not be copyrighted, that artists and media licensing powerhouses might sue AI tech labs for copyrighting infringement, and that governmental institutions might be forcing developers to disclose the datasets they used… going for ethically-sourced data pools seems the way to go. 

Their release of video datasets further proves that AI video generating is the next big thing coming our way, and hopefully, the legal controversies surrounding AI still imagery could be avoided this time around. 

What do you think?

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