me: hey big companies, can I download all of your copyrighted movies and music and images in case one inspires me to make something someday?

big companies: lol no get rekt

also big companies: btw we're downloading and using all of your copyrighted works to stuff into our machine learning models

@april TBF, they did put terms that cover this into that long TOS that you clicked OK without reading.
@mdarweesh @april crawlers who scrape images doesn't have TOS.

@WhyNotZoidberg

Zero of the authors from #book3 and zero of the artists gave permission for their work to be scraped and blended to reproduce slurry.

As you know it also doesn't work because it relies on exploited workers from impoverished countries to provide quality answer models. As you know they're using AI to feed into those models leading to more gibberish.

Nightshade from Ben Zhao at University of Chicago should be on every artist's work.

#aiethics #AIJobs

@mdarweesh @april

@april I look forward to spread the idea of poisoning their machine learning models with misinformation about people, wrong definition, bad quality arts, heinous/outrageous/racist opinions and other garbage ideas.
@ppn @april Fox "News" way ahead of you there
@ppn @april I think you just described the internet
@ppn @april at least their crawlers are getting bad data now; when image scrapers scrape AI images to teach AI it tends to have the opposite effect. Like inbreeding.
@ppn @april: At least one generative AI is trained on 4chan and Stormfront already. They seem to treat racism and fascism as a feature, not a bug.
@april Sure, give us a livable UBI to compensate us all.
@april Exactly. The power of having more cash for lawyers.

@april I argue that the latter weighs even more.

When my copyrighted works are published under BSD/MIT/… licences, they are made available to the average folk with attribution being basically the only “remuneration”. That’s not a lot and easily given by the recipients, so failing to do so weighs worse than failing a part of one of the more complex licences, both copyleft and EULA sides.

ML/LLM/so-called “AI” naturally drops that very attribution by compressing lossily, mixing, and decompressing the inputs. So therefore I’m very much not okay with them using my works under current circumstances.

@april I'm like, Just let me fucking screenshot your video player so I can make silly memes which will inevitably cause more people to actually watch your content.
@april “my lawyer can beat up your lawyer”

@april BTW: "AI" can't create copyrightable works anyway simply because only Authors can create Copyrightable works and only Human Beings can have Authorship!

https://felixreda.eu/2021/07/github-copilot-is-not-infringing-your-copyright/

GitHub Copilot is not infringing your copyright

Felix Reda
lists.d/blocklists.list.tsv at 653d4ef2f70efb4474cf9f31a252974ff4bd037a · greyhat-academy/lists.d

List of useful things. Contribute to greyhat-academy/lists.d development by creating an account on GitHub.

GitHub
@april The worst thing, is that this is the reality. All my data is being kept "hostage" @ Facebook, because i refuse to accept their new terms in Europe, and there's no way i'm paying them to stop doing so either, because i bet we will see a "scandal" in the future, where someone leaks intel about their "dark secret private data profiling". I just don't trust big tech companies like them. I guess i never truly have been lol 😅
@april yeah now more than ever piracy is sooooo legit
@0ddj0bb @april any advice for installing programs that usually require an account? Theoretically
@april Soon, only theaters will have value. Real creativity in real time.
@april
And that's the magic of ✨enclosure✨
@april Unless of course you have granted them permission to do so by e.g. giving your content to Facebook, Instagram, Twitter etc. (Because the ToS, EULA, whatever you clicked away, says so.)
@april big step missing is the authentication/authorization step to view those movies and work which requires agreeing to the terms and conditions. Web scraping is unauthenticated.
@JulianNorton web scraping is governed by robots.txt, and even aside from that an image gallery website has its own terms and services that search engines aren't agreeing to and that includes that they don't have copyright over uploaded images.
@april Correct. Copyright is theft.
@april They won't even let me download their stuff for my OWN language models!

@april not only your copyrighted stuff, also the stuff you thought was private but actually you gave us world-wide irrevocable [etc etc] permission to use it as we want.

also: if your government will require us to publish what we used to train our models we are going to pull out of your market (meaning we will still download your stuff, we just won't sell it back to you remixed with other stuff)...

@april

and this is why people should never use https://doubledouble.top, just think of the shareholders! /s

DoubleDouble - Music Downloads