You’re Not Allowed To Pirate Content But AI Companies Can Easily Afford It

AI-symmetrical Warfare
You may be old enough to recall A&M Records versus Napster or when the Recording Industry Association of America (RIAA) was handing out fines of up to $150,000 per song to kids that had the temerity to pirate their favourite band’s music. The illegality of downloading music was never in question, but the reason that these court cases were so successful was thanks to the fact that the corporations who owned the rights to the artist’s content had vast amounts of money to fund those legal battles and could easily outspend sites like Napster, let alone the individuals they sued.
The same is not true now, as the companies that created and own the LLMs scraping the web of content to train their models can do the exact same thing with little to no consequences.
One company who is a victim of this behaviour, Getty Images, has provided a perfect example of this. They have sued Stability AI, creators of Stable Diffusion, for “spitting out images that replicated Getty’s famous trademark.” The LLM was not just scraping trademarked content, it literally scraped the trademark itself and provided it to users. There is no question about the legality of that action, it is completely and totally illegal. The arguments that Stability AI offered in their defence don’t really matter either, the deciding factor in the ability for LLM developers to happily steal content from artists and corporations that own the rights to content comes down to money.
Getty Images spent millions of dollars on this single court case, but LLMs steal vast amounts of content every day. That means there is no possible way they can afford to defend their rights in court, nor can independent artists even dream of launching a single court case. Thanks to the economic costs content owners face when trying to defend their rights, LLM developers will be allowed to plunder that content regardless of the fact their actions are every bit as illegal as an individual downloading a song. Indeed as the companies are making money from this stolen content it is arguably even more illegal than grabbing song for free for your own personal usage.
Welcome to the future.
According to AI companies, paying artists to train on their works threatens to slow innovation, while rivals in China—who aren't bound by US copyright law—continue scraping the web to advance their models.
More Tech News From Around The Web
- Apple Safari exposes users to fullscreen browser-in-the-middle attacks @ Bleeping Computer
- macOS 26 May Not Support 2018 MacBook Pros, 2019 iMacs, or the iMac Pro @ Slashdot
- Researchers Warn Against Treating AI Outputs as Human-Like Reasoning @ Slashdot
- ASUS to chase business PC market with free AI, or no AI – because nobody knows what to do with it @ The Register
- Threat actors abuse Google Apps Script in evasive phishing attacks @ Bleeping Computer
- Three ways to run Windows apps on a Linux box @ The Register
- Forced E-Waste PCs And The Case Of Windows 11’s Trusted Platform @ Hackaday
- This is the NVIDIA MGX PCIe Switch Board with ConnectX-8 for 8x PCIe GPU Servers @ ServeTheHome
- 3D Print ABS Without A Screaming Hot Bed @ Hackaday