vAIsual, the company that pioneered legally-clean datasets for AI training, has recently announced a first-of-its-kind partnership with the stock photo agency The Middle Frame, which specializes on authentic Middle Eastern and North African imagery.
The goal is to include thousands of The Middle Frame’s inclusive and accurately representative images onto vAIsual's datasets, which other companies can use to train their AI models to produce more diverse, ethical, and legally safe content.
This development aims at improving the quality of content that text-to-image AI tools can create, in an ethical and legal way.
vAIsual was one of the first companies to realize the importance of legal cleanliness in image datasets to train AI generative models. They anticipated the weaknesses in AI-generated images would be precisely in the legal aspect of them – and time would only prove them right, as copyright and legal status of AI-generated content remains the biggest obstacle for its commercial use, and even Getty Images recently sued Stability AI for copyright infringement for using content from their library to train Stable Diffusion.
And so, vAIsual initially launched a unique, legally-clean AI image datasets of people, full of in-house shot portraits with green background, backed by copyright and biometric data releases from models.
Further down the road in 2022, they released DataSetShop.com, an online shop for their datasets, where engineers and researchers can license the different image bundles to train their AI, text-to-image apps.
Now, vAIsual and The Middle Frame's collaboration will bring thousands of the agency's high-quality images, into the DataSetShop's datasets. These images are cleared by the copyright owner, models, and property owners when applicable, to be used in AI software training, thus avoiding all the legal problems of non-authorized image datasets.
The Middle Frame licenses only authentic and modern pictures that accurately represent Arabian culture and people. And this is a key factor in their partnership.
According to vAIsual CEO Michael Osterrieder, “In the past, we have seen huge controversies arise when algorithms depict people from different ethnicities in an inaccurate or offensive way. The best way to circumvent this issue is to make sure the training data includes a diverse range of people.” By introducing thousands of diverse and accurate photos of Middle Eastern communities, they hope to build datasets that will result in more diverse and inclusive, AI-generated images.
This move by vAIsual is a significant one. This company has had the right vision about synthetic media from the get go, and they seem to be on the right track still.
As more entities, organizations and consumers raise doubts about the legal status of AI-generated images, focusing on building legally-safe and visually impactful datasets to train AI software seems to be one of the best ways to give traditional artists and AI creators the right tools to generate content that is ethical and legal to use.
What do you think?
THE AUTHOR
Ivanna Attie
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.
AI Secrets is a platform for tech decision-makers to learn about AI technology. Our team includes experts such as Amos Struck (20+ yrs ICT, Stock Photo, AI), Ivanna Attie (expert in digital comms, design, stock media), and more who share their views on AI.
Get AI news in your inbox & join thousands of engineers and managers using AI to boost sales and grow market share