Research shows that AI image generators can replicate existing photos

Feb 2, 2023

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Research shows that AI image generators can replicate existing photos

Feb 2, 2023

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Join the Discussion

Share on:

Research shows that AI images generators can simply copy an existing photo

It might be a good time to study law. Or copyright law, at least. Researchers have released a study that proves that AI image generators can and do, copy existing images that they’ve ‘looked at’ during the machine learning process. This means that there is a chance that anything spat out by the software could be an exact replica of a copyrighted image.

This debunks the favorite argument that AI machines are no different from humans’ learning processes and that everything they sample is merely ‘inspiration’ to create something new. This appears to not be the case, according to the study. Although it is relatively rare at the moment, the researchers predict that with time it could become a greater problem.

The study was done by researchers at both Google and DeepMind, alongside Berkeley, Cornell, and Princeton universities. The study found that both Google Imagen and Stable Diffusion were capable of directly copying sensitive and copyrighted images used during the training process.

Eric Wallace, a PhD student at UC Berkeley who worked on the study, told Gizmodo that, at the moment, occurrences of image duplication are rare at just 0.3%. However, the research shows that as AI systems become bigger and more powerful, that percentage is likely to increase.

“Maybe in next year, whatever new model comes out that’s a lot bigger and a lot more powerful, then potentially these kinds of memorization risks would be a lot higher than they are now,” Wallace says.

Us humans hold a very low opinion when people copy other people’s work verbatim. But according to this study, the AI is more than capable of doing just that, and it could end up getting into a bit of a pickle over it if it leads to plagiarism on a grand scale, especially as these diffusion-based image generators become more ubiquitous.

We are already seeing some of the stock image giants, such as Getty Images, filing lawsuits against Stable Diffusion et al. If these inconsistencies persist, there may well be grounds for a lot more of them in the future.

[Via Gizmodo]

Filed Under:

Tagged With:

Find this interesting? Share it with your friends!

Alex Baker

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Join the Discussion

DIYP Comment Policy
Be nice, be on-topic, no personal information or flames.

Leave a Reply

Your email address will not be published. Required fields are marked *

2 responses to “Research shows that AI image generators can replicate existing photos”

  1. Naive Guy Avatar
    Naive Guy

    Am I naive concluding that a GAN that you train on one and only one image e. g. “Living in the light with Ann Graham Lotz” will return exactly this image if you prompt it with “Ann Graham Lotz” – and that means any GAN trained on multiple images including “Living in the light
    with Ann Graham Lotz” is theoretically capable of returning exactly this image if you happen to find a precise enough prompt?

    Or put the other way round: Let’s use another GAN guessing prompts which are then further refined by comparing the output of the GAN mentioned above with the target image?

    So the legal implication is: The copyrighted training material is stored somewhere in the parameters of the GAN in a way that it is capable to reproduce exactly that copyrighted material. Without the user even knowing it. And that this behaviour is equal to just reproducing copyrighted material without the consent of the copyright owner.

  2. Martin Gillette Avatar
    Martin Gillette

    So it can do what I can do with a couple key strokes. I’m not impressed.