Data-Scraping Philosophy Tube Puts Ethical AI Into the Pile

CREATOR NEWSLETTER



THE COMMENTS SECTION


So why, besides her recent cameos on Max and Disney shows, am I bringing her up today? Well, Philosophy Tube was one of the YouTube 48,000 channels recently revealed to have been scraped for an illegal, large-scale data dump known as “The Pile.” OpenAI, Apple, Nvidia and Salesforce (among others) are using the Pile to further develop their deep-learning AI products without obtaining consent or offering compensation for the creators whose work is used. OpenAI and Google are now contending with a $5 million dollar lawsuit by a YouTuber on behalf of these content producers, which, as Lon Harris points out, is, uh…nothing. Like, not even a drop in the bucket for the companies, and if split between 170,000 videos that make up the pile’s training set from YouTube would be valuing them at $29 a pop.

Meanwhile, a Meta AI chatbot recently admitted to being trained not on the Pile (yay!), but a wholly separate, proprietary method called Meta Scraping and Extraction (MSaE), which included its own dataset of another 3.7 million YouTube video transcripts. (Boo!) Philosophy Tube was also among those used for MSaE.

This feels particularly unfair, seeing as the pile may very well include Philosophy Tube’s hour-long explainer, “Here’s What Ethical AI Really Means,” which itself is a response to a user taking another popular video on the channel about Effective Altruism and running it through a generative AI program to create non-consensual pornography of Thorn.

“To have someone take my writing and use it for profit – these companies are trying to get people to invest in their product, trying to get people to pay to use their product – and they have taken something that I put a little bit of my heart and soul into, and they are selling it. It makes me feel violated,” Thorn told Marketplace in regards to the pile.

Is there a solution to this thorny issue? I’d suggest supporting Thorn via Patreon and watching her videos on creator-friendly subscription streaming service Nebula. While YouTube has policies forbidding this type of scraping practice, they clearly aren’t taking the sort of action against these major companies that they do among their own users. It’s an irony that won’t escape any of the creators taking financial hits whenever their work is erroneously misidentified by the automated ContentID tool but does nothing to protect their own copyrighted work from being used to help Apple’s profit margins. 


NOTED BY LON HARRIS

Open AI app over youtube logos

IN THE BIZ


PLATFORMS

Twitch’s DJ Program Is Confusing Creators

DJs and streamers have already started to question some of the program’s lop-sided rules, messaging, and rollout.

By Steven Asarch, Passionfruit Contributor


TIPS & TRICKS

How To Make A Sound on TikTok: A Guide To Working The System

Get control over how you use sound on TikTok.

By Jen Glantz, Passionfruit Contributor


WHAT WE’RE WATCHING


YOUTUBE MADE ME DO IT

Let’s get weird with Eliot Dewberry, co-host of Internet Today, as he joins Deep Linkers to discuss his journey from ETC and Machinima to co-running his own channel with co-host Ricky Hayberg. They now explore the weird, wild world of internet culture, covering everything important, troublesome, funny, and bizarre.

Be sure to subscribe to the Passionfruit YouTube channel so you don’t miss an episode! If you’d prefer to listen in audio form, we also have a podcast feed

Content for Creators.

News, tips, and tricks delivered to your inbox twice a week.

Newsletter Signup

Latest Newsletters

  • 🤖 Creators Speak Out Against AI Art

    🤖 Creators Speak Out Against AI Art

    CREATOR ECONOMY NEWSLETTER Issue #93 | Dec. 20, 2022 This month, discourse around artificial intelligence (AI) art picked up traction once again since the technology’s emergence years ago, after viral filters and software like MyHeritage AI and Lensa blew up on TikTok. TikTokers are infatuated with how MyHeritage showcases “you” from different historical periods using…

  • 🪦 Twitter Isn’t Dead—But Are Its Competitors?

    🪦 Twitter Isn’t Dead—But Are Its Competitors?

    CREATOR ECONOMY NEWSLETTER Issue #92 | Dec. 15, 2022 Over a month since Elon Musk’s takeover of Twitter, the platform still seems to be in chaos. Just this week, the company caught heat for disbanding its Trust and Safety council. It then confused creators with the launch of a gold and blue verification checkmark system.…

  • 💰 What Can We Learn From YouTube’s All-Time Top Earners?

    💰 What Can We Learn From YouTube’s All-Time Top Earners?

    CREATOR ECONOMY NEWSLETTER Issue #91 | Dec. 13, 2022 Top-earning YouTubers rake in millions of dollars each year from brand deals, merchandising, and launching product lines.  While TikTok won the attention of many Gen Zers, the top earner on TikTok in 2021, Charli D’Amelio, only earned $17.5 million. That’s less than what the top nine…