#AI#GenerativeAI#MidJourney#StabilityAI#GeneratedImages#WebScraping: "On Wednesday, Midjourney banned all employees from image synthesis rival Stability AI from its service indefinitely after it detected "botnet-like" activity suspected to be a Stability employee attempting to scrape prompt and image pairs in bulk. Midjourney advocate Nick St. Pierre tweeted about the announcement, which came via Midjourney's official Discord channel.
Prompts are the written instructions (like "a cat in a car holding a can of a beer") used by generative AI models such as Midjourney and Stability AI's Stable Diffusion 3 (SD3) to synthesize images. Having prompt and image pairs could potentially help the training or fine-tuning of a rival AI image generator model."
#UK#AI#GenerativeAI#GeneratedImages#StabilityAI#GettyImages#Copyright: "The core of the claimants’ allegations is that Stability AI scraped millions of images from the Getty website without consent, and used these images to train and develop Stable Diffusion. It’s further claimed that the synthetic images generated by Stable Diffusion, accessed by users in the UK, infringe upon Getty Images’ copyrighted works and bear their trade marks. Some of these images had been presented in the particulars of claim, but it was never made clear how the
images came to be. I was able to produce some images myself with older versions of Stable Diffusion bearing the semblance of a Getty Images logo, but none of the outputs produced appeared to come from the images in the input. The idea here is that Stable Diffusion “memorised” the Getty logo, and could place it on outputs on demand. This is no longer possible as far as I can tell."
I'm at the Intel Innovation conference today in San Jose. I expect a coming-out party for Meteor Lake, Intel's next-gen PC processor arriving later this year. Plus lots of chatter about chipmaking, AI, Xeon, development software. It wouldn't be an Intel event without lots of 300mm wafers. 1/n #Intel#processors#semiconductors#chips
Gelsinger shows off his coding skills (on his T-shirt) to make the case that AI coding is easier. Stability.ai will buy 4,000 Intel Gaudi AI processors, overseen by Xeon server chips, for what he says will be the largest AI machine in Europe. 5/n #Intel#AI#stabilityai
Stability AI released the StableCode - their first LLM generative AI product for coding today. The new model - StableCode-Completion-Alpha-3B is a 3 billion parameter decoder-only code completion model pre-trained on a diverse set of programming languages using Stackoverflow developer survey data.
🚫 Getty asks London court to stop UK sales of Stability AI system
➥ Reuters
"The Seattle-based company accuses the company of breaching its copyright by using its images to "train" its Stable Diffusion system, according to the filing dated May 12."
The enforcement of copyright law is really simple.
If you were a kid who used Napster in the early 2000s to download the latest album by The Offspring or Destiny's Child, because you couldn't afford the CD, then you need to go to court! And potentially face criminal sanctions or punitive damages to the RIAA for each song you download, because you're an evil pirate! You wouldn't steal a car! Creators must be paid!
If you created educational videos on YouTube in the 2010s, and featured a video or audio clip, then even if it's fair use, and even if it's used to make a legitimate point, you're getting demonetised. That's assuming your videos don't disappear or get shadow banned or your account isn't shut entirely. Oh, and good luck finding your way through YouTube's convoluted DMCA process! All creators are equal in deserving pay, but some are more equal than others!
And if you're a corporation with a market capitalisation of US$1.5 trillion (Google/Alphabet) or US$2.3 billion (Microsoft), then you can freely use everyone's intellectual property to train your generative AI bots. Suddenly creators don't deserve to be paid a cent.
Apparently, an individual downloading a single file is like stealing a car. But a trillion-dollar corporation stealing every car is just good business.
#Copyright does not protect the concept and themes of artistic presentation. So training autocomplete tools like #ChatGPT or generative art tools along the lines of #StabilityAI on huge amounts of copyrighted material available on the web doesn't seem to trespass on the rights actually created by copyright law. That is, neither the trained model parameters nor the output qualifies as a infringing copy.
The fact that big corporations have heated the rhetoric with even small-scale copyright infringement being characterized as if it were an existential threat rather than free marketing perhaps misleads people to think copyright grants the owner total control of the future of their creations. But law is about statutes and precedents, not feelings, which is why big corporations aren't likely to train their models on billions of copyrighted works if there was a credible risk of paying statutory damages on a per-work basis.
If there is a moral right to the "something" that has been gifted to these models by their training, it has not been well described, let alone recognized in law as property of the creators. How is this "something" which an AI model steals supposed to be distinguished from the piecewise appreciation for the art as summed over all human viewers?
So perhaps the real problem is the moral outrage created by the corporations who for decades equate copyright infringement with being ambushed by a gang of seagoing rapists, kidnappers, killers, and robbers (pirates). Towards that end, Germany is discussing adding copyright infringement as a form of "digital violence" making the analogy more exact.