If you want to know how AI companies are trying to destroy creators' livelihoods, you *have to scrape the AI companies*. You can't ask companies for permission to gather information that might be used to destroy them - they'll just say no. If taking information off the internet without permission is "theft," then gathering information by scraping AI companies is *also* theft.

37/

"Unpermissioned research" isn't a free-for-all. Universities have ethical rules designed to protect the privacy rights and other human rights of research subjects, and because these aren't property rights, they can be balanced against the socially beneficial outcomes of research. Universities can get this wrong, of course, but when they do, it's not *theft*. It's a human rights violation, a privacy violation, a labor violation.

36/

"Unpermissioned research" seeks to circumvent limits that platforms establish *specifically* to stop outsiders from learning how they operate. When you're doing unpermissioned research, you try to get around rate limits, query throttles, and other measures that platforms use to block others from mapping their extent and documenting their conduct.

35/

Zuckerman studies platform power: "using data to answer hard questions about social media, search engines and AI tools." He leads a team that is documenting exactly, precisely how tech companies collude with authoritarians to spy on us, manipulate us, and control us. And his methodology is something called "unpermissioned research," which is what academics call scraping:

https://www.techpolicy.press/ai-companies-threaten-independent-social-media-research/

34/

The people who keep track of this stuff *are* worried about being fined or arrested. Ethan Zuckerman, one of America's foundational internet scholars, has just accepted Canadian government funding to move his lab from UMass to McGill in Montreal:

https://ethanzuckerman.com/2026/08/27/my-personal-contribution-to-the-us-canada-trade-war/

33/

Sure, if you do this carelessly or recklessly, you can end up violating someone's labor rights, or privacy rights, or human rights. And because *those* frameworks *aren't* based on the sanctity of property rights, *they* can be used to protect these important rights *without* giving corporate America the right to have you fined or arrested for documenting their takeover of the America.

32/

Same goes for keeping track of genocide apologists, data-center astroturfers and ICE cheerleaders who've flooded Tiktok ever since Trump stole it and handed it over to his creepy billionaire pal Larry Ellison.

Making copies of that stuff isn't theft. It's not a copyright violation. Not even if you do it to billions of works. Not even if it's bad for the companies whose feeds you're capturing. Not even if it's bad for the dark money groups who funded the content.

31/

Scraping is how we produce evidence of the changes that powerful people are making to the world around us. Do you want to know whether Mark Zuckerberg or Elon Musk are downranking content critical of Trump and Big Tech and pumping racist and conspiratorial posts into the resulting void? You'd better hope you can scrape the feeds they cram into billions of people's eyeballs.

30/

Remember our earlier discussion of how privacy violations are weaponized to make poor people even poorer, by depressing their wages and raising prices based on inferences about their economic desperation? Our best weapon for fighting this practice is *scraping*, because that's how we catch corporations changing prices and wages based on surveillance data:

https://pluralistic.net/2023/09/17/how-to-think-about-scraping/

29/

There's exactly *one* company that could pull that off: Google, a rapacious tech monopolist that is - not coincidentally - one of the leaders of the movement to beggar every creative worker. We will not improve the world, the internet, or creative workers' lives by ensuring that the last search engine anyone ever creates is Google.
28/