Just today I noticed a startling number of old-age versions of friends on Facebook, at wonder with how creepily true they seem to be. The AI-generated photos are provided via FaceApp, which seems both creepy and cool at the same time.
So how does a free service put all their AI (artificial intelligence) hardware to use in the background and not go bankrupt? In other words, what does “free” really cost you, and how does “free” pay for all the expensive hardware to run it? This is the question we should all ask ourselves of any “free” service, whether it’s FaceApp, Facebook, Youtube, Android, Google, Twitter, etc. And unfortunately, the answer isn’t always so easy to figure out. Youtube runs ads, but what does FaceApp get out of it?
The first thing they get is a clean, high-resolution photo of the real you. So what? Well, that’s the only input they receive from you, so apparently that’s what’s paying their bills. Did you know your photo was worth that much?
As seen in FaceApp’s Terms Page:
“You grant FaceApp a perpetual, irrevocable, nonexclusive, royalty-free, worldwide, fully-paid, transferable sub-licensable license to use, reproduce, modify, adapt, publish, translate, create derivative works from, distribute, publicly perform and display your User Content and any name, username or likeness provided in connection with your User Content in all media formats and channels now known or later developed, without compensation to you.”
Excerpt from FaceApp Terms page, 17 July 2019
Wow, that’s pretty inclusive of just about any imaginable way they could possibly monetize your image, right?
You may not care, and that’s fine. However, there are some troubling technologies out there that can make excellent use of your image. In fact, “deepfake” technology can graft your image (quite convincingly) into videos of other people saying or doing other things, such as this grafting of Steve Buscemi’s face on the body of Jennifer Lawrence during an interview, as tweeted by Mikael Thalen.
Snapshot from Mikael Thalen’s tweeted video
And judging by the high percentage of porn results when searching for “deepfake”, misuse and monetization of your image is readily available to the highest bidder (or is that the lowest?).
And this is just one of the technologies we know about now. What about those “later developed” or otherwise sub-licensed to third parties? Though fluid deepfake video currently requires more than one still image of any given person, coming technological improvements may eliminate that requirement in the near future.
Is there really any limit to this? Are we facing a world in which the difference between reality and fiction will one day become indistinguishable? Are we already there?
If you’d like to learn more about this topic, here are some articles I’ve found useful:
FaceApp: Privacy warning issued over app that lets users turn themselves old
Samsung deepfake AI could fabricate a video of you from a single profile pic
Deepfakes may ruin the world. And they can come for you, too