I'm surprised they're doing this, because it could open them up to legal liability. If they have the technology to do this, then arguably they have the technology to detect child abuse or other crimes and would be liable for hosting it. If they didn't have this technology, then if they are accused of hosting illegal stuff, they can just shrug and say there's nothing they can do. Or, like you said, remove individual items after a complaint.

This is a myth.

Section 230 of the Commucations Deceny Act of 1996 "allows ISPs and other service providers to restrict customers' actions without fear of being found legally liable for the actions that are allowed."

"The act was passed in part in reaction to a 1995 decision which suggested that service providers who assumed an editorial role with regard to customer content thus became publishers, and legally responsible for libel and other torts committed by customers. This act was passed to specifically enhance service providers' ability to delete or otherwise monitor content without themselves becoming publishers"

Source: https://en.wikipedia.org/wiki/Section_230_of_the_Communicati...

Section 230 appears to be changing.

https://www.npr.org/sections/alltechconsidered/2018/03/21/59...

https://www.npr.org/sections/thetwo-way/2018/03/23/596460672...

Yes and no: SESTA/FOSTA make business liable for certain criminal content on their platforms. But the Section 230 shield wasn't about liability in general, but rather about protecting you from an increase in liability as a result of moderation.

To use an analogy: Sec 230 is like a "Good Samaritan" law that protects people who try to help accident victims from liability for their actions. This is obviously good public policy, because many people are scared of being sued if they make any mistakes (or even if they don't), even though the evidence points overwhelmingly to a net positive effect of such attempts of help.

Backpage used Section 230 like a guy punching someone in the face and then claiming they were trying to protect them from the mosquito sitting on their nose.

The new law assigns liability if you are grossly negligent in allowing certain content to flourish on your platform.

But, the important point is: this liability is independent of any moderation platforms may or may not engage in. The attempt to moderate some content still does not increase your liability.

Check out jepZ's link above: https://news.ycombinator.com/item?id=16668874

Safe harbor is effectively dead. Recent changes in law make neutral content hosters liable for the content their users create/upload/share (going beyond the standard strictly illegal stuff like child porn, inciting violence, etc).

My guess is that it opens them up to legal liability not to: https://en.wikipedia.org/wiki/FOSTA

(I work for Google, but I don't have any inside information on this; it's just speculation.)

They do detect child abuse imagery and report it to the appropriate agencies.

It is more about that new law than about their ToS:

https://news.ycombinator.com/item?id=16655494

Thats pure speculation.

Yes, you are right, I didn't see that the article was from before the vote and didn't read it carefully enough.

It could be similar to DropBox's sharing copyrighted material ban, from what I vaguely recall reading somewhere, it merely compares the hash with what's known to be pirated like torrent trackers, though if I'm wrong would be happy to be corrected.

The content that's being discussed in the article is being created by the uploader. If it's being detected by an automated system, it's a disturbingly intelligent one.

Google is however nowadays staffed by activists.