Would applying a watermark to all the training images force the AI to add a watermark?
Comment on China will enforce clear flagging of all AI generated content starting from September
riskable@programming.dev 5 weeks ago
Not a bad law if applied to companies and public figures. Complete wishful thinking if applied to individuals.
For companies it’s actually enforceable but for individuals it’s basically impossible and even if you do catch someone uploading AI-generated stuff: Who cares. It’s the intent that matters when it comes to individuals.
Were they trying to besmirch someone’s reputation by uploading false images of that person in situations they were never in? That’s clear bad intent.
Were they trying to incite a riot or intentionally spreading disinformation? Again, clear bad intent.
Were they showing off something cool they made with AI generation? It is of no consequence and should be treated as such.
RandomVideos@programming.dev 5 weeks ago
riskable@programming.dev 5 weeks ago
Nope. In fact, if you generate a lot of images with AI you’ll sometimes notice something resembling a watermark in the output. Demonstrating that the images used to train the model did indeed have watermarks.
Removing such imaginary watermarks is trivial in image2image tools though (it’s just a quick extra step after generation).
RandomVideos@programming.dev 5 weeks ago
I meant having all training images with a watermark, not only some of them
lily33@lemm.ee 5 weeks ago
I agree that it’s difficult to enforce such a requirement on individuals. That said, I don’t agree that nobody cares for the content they post. If they have “something cool they made with AI generation” - then it’s not a big deal to have to mark it as AI-generated.
riskable@programming.dev 5 weeks ago
Why stop at “AI-generated”? Why not have the individual post their entire workflow, showing which model they used, the prompt, and any follow-up editing or post-processing they did to the image?
In the 90s we went through this same shit with legislators trying to ban photoshopped images (hah: They still try this from time to time). Then there were attempts at legislating mandatory watermarks and similar concepts. It’s all the same concept: New technology scary, regulate and restrict it.
In a few years AI-generated content will be as common as photoshopped images and no one will bat an eye because it’ll “just be normal”. A photographer might take a picture of a model (or a number of them) for a cover or something then they’ll use AI to change the image after. Or they’ll use AI to generate an image from scratch and then have models try to copy it. Or they’ll just use AI to change small details in the image such as improving lighting conditions or changing eye color.
AI is very rapidly becoming just another tool in photo/video editing and soon it will be just another tool in document writing and audio recording/music creation.
Buelldozer@lemmy.today 5 weeks ago
We’re already there you just aren’t noticing them anymore because they’ve progressed beyond the Six Fingers / Halo Ring in the eyes level of believability.
Stanley_Pain@lemmy.dbzer0.com 5 weeks ago
This really underscores the need for complete reform of the entire media apparatus…