How does Lensa AI work? Why do people have serious concerns
Lensa AI, Stable Diffusion, and TrueDepth Technology
It’s taken social media by storm. No doubt you’ve seen the incredibly vibrant and detailed artwork of your friends, flooding your feeds. Millions of people downloaded the app in the first week it became available, and just by looking at the eye-catching images, it’s no surprise why this new technology has gained so much popularity in such a short period of time.
Lensa is a product of the US based Prisma Labs, and the pictures you’ve seen on social media are a result of the “Magic Avatar” functionality inside the app.
At its core, the app’s artificial intelligence system employs a recombinatory approach known as “Stable Diffusion.” Scraping billions of images across the internet to train algorithm’s neural network.
This data set is then combined with TrueDepth API technology, which analyzes the user-provided photos ("face data"), projecting thousands of points on to the face and creating a “depth map” to then, according to the developers, “train our algorithms to perform better and show you better results."
You may already be familiar with this technology, since it is the engine behind Face ID, the system that detects, recognizes and matches your face to the registered facial data and unlocks your Apple device.
Digital artists, intellectual property, and privacy concerns
Is there a price to paid for this technology? Even though the digital artwork which the artificial intelligence is trained on is publicly available on the internet, many artists say that these images are copyrighted and that the app is allegedly violating their copyright by using their intellectual property without permission, credit, or compensation.
Furthermore, for the user of the app, there are concerns about the ownership of this image and what that means for your privacy:
The perpetual ownership of this information has many people concerned about data security.
Although the terms and conditions say that users “retain all rights in and to your user content," using the app gives them “perpetual, revocable, nonexclusive, royalty-free, worldwide, fully-paid, transferable, sub-licensable license to use, reproduce, modify, adapt, translate, create derivative works.”
To assuage users, the company has recently stated on Twitter that “As soon as the avatars are generated, the user’s photos and the associated model are erased permanently from our servers."
What are your thoughts about this new extremely popular visual technology?
