Meta drops the word "train" from its new terms, but lawyers say the scraping loophole stays wide open
Legal experts warn Meta's new terms still grant broad scraping permissions
Meta revised its upcoming business terms after Nieman Lab revealed they would let it scrape sites "associated with" business accounts. The word "train" is gone, and permitted uses are now limited to "advertising and business products." But legal experts say "develop" is broad enough to cover training AI, and nothing restricts scraping whole sites. Meta insists robots.txt preferences still apply, though the terms don't say so. The new rules take effect October 30, with no opt-out.
While it adds more language about the permitted uses, "develop…advertising and business products" is still broad enough to drive a truck through. This does nothing to restrict the "associated with your products or services" language that would let Meta scrape entire sites.
- Bender
still have the ability to disable the crawler via robots.txt flags
Never rely on robots.txt. Configure it, but do not rely on it. A bot will do what a bot can do. Their bots are trivial to spot so close the connection, bonus if done silently closing the socket and inhibiting first RST from web daemon and second RST from the OS.
- Retro_Dev
I'm a bit confused why people are freaking out about these term changes - Meta scrapes my website already on a fairly regular basis (even with robots.txt). Is the point of this change to justify more intense scraping?
- asanineassasin
The valuable data to scrape is in the IDE of the expert and that expert getting challenges.
- newhotelowner
And they spend huge resource so that we can't scrap Facebook - our own feed.
- jareklupinski
just rename the company to Meta Law or Zuckerberg & Associates already