The pioneer in legally clean AI training datasets, vAIsual, recently announced a new first-of-its-kind partnership with The Middle Frame, a stock photo site focused on Middle Eastern and North African imagery.
This deal will bring The Middle Frame’s authentic and inclusive images onto datasets that other companies can use to train their AI models to produce more diverse, ethical, and legally sound results.
Here we’ll share all the details of this interesting development!
The newly announced partnership between vAIsual and The Middle Frame means that thousands of realistic images that accurately represent Arabian culture and people will be added to datasets on DataSetShop.com; this is vAIsual’s online shop for datasets, where engineers and researchers can license different image datasets to train their AI, text-to-image apps that will later produce synthetic media.
vAIsual was one of the first, if not the first, companies to realize the imperious need for legally-clean image datasets to train AI generative models, anticipating the legal mud that AI-generated images would be surrounded by. They launched the first-ever, legally-clean AI image datasets of people, produced entirely in-house and backed by copyright and biometric data releases from models at the end of 2021.
Time would prove them right, as in 2022, many questions regarding the legal status of AI-generated images surged, from who owns the copyright of images created by an AI tool, to whether it’s ethical to use other artists’ work to train AI software without their permission. And more recently, in 2023, Getty Images has gone out with a big lawsuit against Stability AI for the use of their content in the training of Stable Diffusion. Forbes has stated that 2023 will be the year of AI ethics legislation, so much development is expected in this field.
vAIsual’s datasets avoid all those legal problems altogether by providing image bundles that are cleared by the copyright owner, models, and property owners when applicable.
If you train your software with their datasets –or use tools that have been trained with them– you are assured that your AI-generated pictures will be ethically and legally safe.
Besides the legal aspect, vAIsual aims at solving another pressing problem that AI-created visuals have been facing: inaccurate representation.
According to vAIsual’s 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.”
Their teaming up with The Middle Frame seeks to provide engineers with diverse and authentic representations of different ethnicities and cultures so that their apps can produce more faithful and inclusive content.
For The Middle Frame, this deal is a way to enter the fast-growing industry of AI visuals on the right foot. According to co-founder and COO Mohammad Alnobani, they “want to be at the forefront of how the stock industry participates in the future of IP licensing, and this deal is an ideal way to start.”
While vAIsual and The Middle Frame aren’t the only companies joining forces to tackle the obstacles in AI-generated images for the stock licensing industry, their collaboration is certainly unique in the sense that it pursues visual diversity and not only legal safety.
And to that, we have to add the DataSetShop offer that puts these curated datasets at the reach of any developer rather than for internal use only.
Overall, one thing is for sure, and that is that, so far, vAIsual is still at the vanguard of the industry in terms of vision and steps forward.
We are excited to see what new collaborations come up and what the first text-to-image tools trained with DataSetShop bundles can produce!