Getty Images Sues AI Art Generator Stable Diffusion in the US For Copyright Infringement
Read more of this story at Slashdot.
Computers Tech Games Crypto Music and More
Read more of this story at Slashdot.
However, Carlini’s results are not as clear-cut as they may first appear. Discovering instances of memorization in Stable Diffusion required 175 million image generations for testing and preexisting knowledge of trained images. Researchers only extracted 94 direct matches and 109 perceptual near-matches out of 350,000 high-probability-of-memorization images they tested (a set of known duplicates in the 160 million-image dataset used to train Stable Diffusion), resulting in a roughly 0.03 percent memorization rate in this particular scenario. Also, the researchers note that the “memorization” they’ve discovered is approximate since the AI model cannot produce identical byte-for-byte copies of the training images. By definition, Stable Diffusion cannot memorize large amounts of data because the size of the 160,000 million-image training dataset is many orders of magnitude larger than the 2GB Stable Diffusion AI model. That means any memorization that exists in the model is small, rare, and very difficult to accidentally extract.
Still, even when present in very small quantities, the paper appears to show that approximate memorization in latent diffusion models does exist, and that could have implications for data privacy and copyright. The results may one day affect potential image synthesis regulation if the AI models become considered “lossy databases” that can reproduce training data, as one AI pundit speculated. Although considering the 0.03 percent hit rate, they would have to be considered very, very lossy databases — perhaps to a statistically insignificant degree. […] Eric Wallace, one of the paper’s authors, shared some personal thoughts on the research in a Twitter thread. As stated in the paper, he suggested that AI model-makers should de-duplicate their data to reduce memorization. He also noted that Stable Diffusion’s model is small relative to its training set, so larger diffusion models are likely to memorize more. And he advised against applying today’s diffusion models to privacy-sensitive domains like medical imagery.
Read more of this story at Slashdot.
Last September Getty Images banned the inclusion of AI-generated works in its commercial database over copyright concerns. On Tuesday, Getty Images announced that it is suing Stability AI, maker of the popular AI art tool Stable Diffusion, in a London court over alleged copyright violations.
“It is Getty Images’ position that Stability AI unlawfully copied and processed millions of images protected by copyright and the associated metadata owned or represented by Getty Images absent a license to benefit Stability AI’s commercial interests and to the detriment of the content creators,” Getty Images wrote in a press statement released Tuesday. “Getty Images believes artificial intelligence has the potential to stimulate creative endeavors.”
“Getty Images provided licenses to leading technology innovators for purposes related to training artificial intelligence systems in a manner that respects personal and intellectual property rights,” the company continued. “Stability AI did not seek any such license from Getty Images and instead, we believe, chose to ignore viable licensing options and long‑standing legal protections in pursuit of their stand‑alone commercial interests.”
The details of the lawsuit have not been made public, though Getty Images CEO Craig Peters told The Verge, that charges would include copyright and site TOS violations like web scraping. Furthermore, Peters explained that the company is not seeking monetary damages in this case so as much as it is hoping to establish a favorable precedent for future litigation.
Text-to-image generation tools like Stable Diffusion, Dall-E and Midjourney don’t create the artwork that they produce in the same way people do — there is no imagination from which these ideas can spring forth. Like other generative AI, these tools are trained to do what they do using massive databases of annotated images — think, hundreds of thousands of frog pictures labelled “frog” used to teach a computer algorithm what a frog looks like.
And why go through the trouble of assembling and annotating a database of your own when there’s an entire internet’s worth of content there for the taking? AI firms like Clearview and Voyager Labs have already tried and been massively, repeatedly fined for scraping image data from the public web and social media sites. An independent study conducted last August concluded that a notable portion of Stable Diffusion’s data was likely pulled directly from the Getty Images site, in part as evidenced by the art tool’s habit of recreating the Getty watermark.
The group trying to monetize AI porn generation, Unstable Diffusion, raised more than $56,000 on Kickstarter from 867 backers. Now, as Kickstarter changes its thinking about what kind of AI-based projects it will allow, the crowdfunding platform has shut down Unstable Diffusion’s campaign. Since Kickstarter runs an all-or-nothing model and the campaign had not yet […]
Kickstarter shut down the campaign for AI porn group Unstable Diffusion amid changing guidelines by Amanda Silberling originally published on TechCrunch
When Stable Diffusion, the text-to-image AI developed by startup Stability AI, was open sourced earlier this year, it didn’t take long for the internet to wield it for porn-creating purposes. Communities across Reddit and 4chan tapped the AI system to generate realistic and anime-style images of nude characters, mostly women, as well as non-consensual fake […]
Meet Unstable Diffusion, the group trying to monetize AI porn generators by Kyle Wiggers originally published on TechCrunch