Hundreds of thousands of YouTube videos have been used by successful tech companies to train AI models -- which might not have been a big deal if the creators of those videos had given them permission to do so.
WIRED recently reported that over 100,000 YouTube videos -- many from well-known creators such as MrBeast -- were scraped for AI training purposes. The owners of such channels say they never gave consent for this activity.
“No one came to me and said, ‘We would like to use this,’” The David Pakman Show host David Pakman told Proof News. Over 100 of his videos were used to train AI -- and Apple isn't even the only company allegedly responsible.
Tech companies such as Salesforce and Nvidia were also among the reported scrapers, potentially violating You'Tube's rules against mining user content without consent. According to YouTube CEO Neal Mohan, using such content specifically for AI training is against the platform's terms and conditions.
So the question becomes: Will companies be able to get away with scraping user content across the internet to train its AI models? It depends on the result of lawsuits like the one The New York Times filed claiming copyright violations. It may be up to large companies demanding regulations around the use of AI and its training models to drive change, but content creators can also continue to speak up about their concerns.
According to Proof News's findings, the data collected from these YouTube videos for AI training purposes was mostly text-based. On Reddit, some users speculate that companies might be gathering their data from video captions.
The use of AI has divided people in the online space, especially creatives who understand their work is likely being used to train AI and they aren't being compensated for it. Many don't seem to mind, but others are demanding more regulations to stop AI from stealing creators' work.
How the digital space continues to evolve as companies invest heavy resources into AI training remains uncertain.