Can Google protect against copyright theft by Generative AI?
As someone who creates a lot of images (have you seen my Tumblr?) I’m obviously concerned about stuff getting nicked. Much of my work gets watermarked, but scraping it off is easy. Adobe Photoshop even ships with all you need to do it.
Google has come up with an idea: to embed a watermark deep into the image’s code. Yes, images are also code. Each JPEG is a set of numbers that tell your computer what color each pixel is.
The theory is an embedded watermark can escape the filtering, flipping, recoloring and other tricks image thieves apply to take credit for your work. Or “repurpose” it for evil.
Obviously big brands will love it, but so should small-scale creators making a nickel off their latest creation. You might not be able to stop it being pinched, but it could make it easier to track these images down and even pursue the miscreants if that’s your thing.
Of course, this is likely to trigger another technological battlefield. Google’s work today will undoubtedly be undone by someone tomorrow, and around in circles we go. Caught in the middle of this are the creators struggling to make a few yen off the back of their art and photography.
I’m watching this for now and will see how it plays out over the coming months. It’s on a small-scale trial at the moment, although it can only be a matter of time before Google pushes it out to a wider audience. Given their love of adding centralized tracking to everything they touch, we creators will have to decide what’s more important…
protecting our copyright or other people’s privacy?