4 ms·

My recollection of history is that 230 popped out of the need to protect ISPs (which were originally smaller businesses) from someone doing something illegal th

My recollection of history is that 230 popped out of the need to protect ISPs (which were originally smaller businesses) from someone doing something illegal through their servers. Until then, if I recall correctly (and it's entirely probable I don't), then even things like forfeiture laws could come into play. You're a mom and pop ISP, you host some websites on pizza box Sparc servers, and suddenly the FBI is scooping it up while busting and internet crime ring. I don't think Google should be held liable if I put up something illegal. But I don't see how you can argue the algorithm is entirely neutral, if the algorithm is to drive engagement. The algorithm is not recommending a random video with 'cooking' in its description. It's going from 'cooking', to 'eat more vegetables', to 'have you considered vegetarianism,' eventually to 'vegan purity must be maintained - meat eaters must die' and 'we gloriously behead a poultry farmer.' At some point, when you're making decisions about what to recommend, you're no longer just passively hosting content. Even if the algorithm wasn't intentionally written to turn people into a radical vegetablists, at some point it shouldn't matter. And (again, my recollection) Google didn't jump up and say "Gadzooks! You're right! We shall cease immediately for we are not evil (despite our handle-bar mustaches)" They, and other social content companies, fought the changes. So it's not as if they were ignorant of what they were doing. Meh - I think some of the other posters are right - they'll kick it down the road and hope it goes away.

krapp 4y ago
>But I don't see how you can argue the algorithm is entirely neutral, if the algorithm is to drive engagement. Section 230 doesn't require that platforms be neutral.
- phoehne 4y ago
  I was thinking of the quote in the article from Justice Thomas, more than a statement about the political or ideological neutrality of the platform: > “If it’s the same algorithm to present cooking videos to people who are interested in cooking and ISIS videos to people who are interested in ISIS, racing videos to people who are interested in racing, I think you’re going to have to explain more clearly if it’s neutral in that way how your claim is set apart from that,” Thomas told Schnapper. I don't think it's just a matter of representing racing videos to people interested in racing. For example, I start with a video on police misconduct and before I realize it, I'm seeing videos telling me as a sovereign citizen I can can't be arrested ... because magic.
joshuamorton 4y ago
You're misremembering. 230 didn't really ever regard ISPs, it applied to what were functional web forums, CompuServe which didn't moderate and Prodigy which did. CompuServe didn't moderate at all, while Prodigy did moderate. The courts found Prodigy liable for content, because it chose to moderate some content. CompuServe was not liable because it left all content up. Congress wanted to encourage sites to moderate content without suddenly becoming liable for anything they failed to moderate, so 230 exists to shield sites who choose to moderate some content from becoming liable for all the other content that they fail to moderate in good faith (https://www.eff.org/issues/cda230/legislative-history https://www.eff.org/issues/cda230/legislative-history).
- phoehne 4y ago
  Thanks. The crux of this is just about what they failed to moderate and if they were acting in good faith?
  - jcranmer 4y ago
    The crux is that the plaintiffs want to go after Google in any way they can, and they're grasping at straws to try to eviscerate §230 immunity. (No really, they're arguing that the URLs in search recommendations are original content that therefore doesn't get §230 immunity). §230 is explicitly pretty broad on immunity: it states that providers are neither liable for failing to moderate nor for the consequences of their moderation. "Good faith" is a requirement, but only for actions taken to restrict access to the material (i.e., it doesn't apply to a failure to moderate.)
ajross 4y ago
Could you maybe list some examples less contrived than "we gloriously behead a poultry farmer"? Needless to say completely fictitious partisan warfare is unlikely to convince an actual court. In point of fact there probably is some stuff to worry about, but where we've seen it it's been almost exclusively in off label sites. Unpopular opinion: the core tech portals have been remarkably good stewards of the public good in the face of genuine extremis in the discourse the public wants to engage in. Everyone's got their favorite Moment of Clear Censorship that Must Not Be Permitted, but name any forum and any topic, and I'll be able to show you (with a quick seearch using tools provided by these very tech giants) extensive unimpeded discussion. To wit: SCOTUS isn't going to step in because one tweet got taken down. There are thousands just like it that weren't censored.
- phoehne 4y ago
  I'm torn on that. I finally just read the actual text instead of relying on my memory of 1996. I get what you're saying, but it isn't a discussion like this one here. It's more like the platforms pick and choose what you're likely to see next. There's not just showing up in the order they were posted or the most relevant things. It might be by how angry they predict it will make me, continue to engage, and get more impressions. You see it in crowds, for example, when people go from having one energetic but good time to being put in a state where they are likely to commit violence. I don't know if Travis Scott is more legend than phenomena, or if the Woodstock '99 people realized they created a metal festival. Human beings are weird creatures and we can do horrific things, given the right context and environment. I think the actual text is rarely a cause for concern, but the manipulation around it is. You can take the same group of people and have an orderly discussion on defenestration, but work them up and you can get an actual defenestration. So if a company is manipulating the flow of information to create an emotional state, on purpose, I'm not sure that company should be 100% in the clear on it. Especially since the law seems to be more about good faith efforts at restricting access to content.

krapp 4y ago

>But I don't see how you can argue the algorithm is entirely neutral, if the algorithm is to drive engagement. Section 230 doesn't require that platforms be neutral.

phoehne 4y ago
I was thinking of the quote in the article from Justice Thomas, more than a statement about the political or ideological neutrality of the platform: > “If it’s the same algorithm to present cooking videos to people who are interested in cooking and ISIS videos to people who are interested in ISIS, racing videos to people who are interested in racing, I think you’re going to have to explain more clearly if it’s neutral in that way how your claim is set apart from that,” Thomas told Schnapper. I don't think it's just a matter of representing racing videos to people interested in racing. For example, I start with a video on police misconduct and before I realize it, I'm seeing videos telling me as a sovereign citizen I can can't be arrested ... because magic.

joshuamorton 4y ago

You're misremembering. 230 didn't really ever regard ISPs, it applied to what were functional web forums, CompuServe which didn't moderate and Prodigy which did. CompuServe didn't moderate at all, while Prodigy did moderate. The courts found Prodigy liable for content, because it chose to moderate some content. CompuServe was not liable because it left all content up. Congress wanted to encourage sites to moderate content without suddenly becoming liable for anything they failed to moderate, so 230 exists to shield sites who choose to moderate some content from becoming liable for all the other content that they fail to moderate in good faith (https://www.eff.org/issues/cda230/legislative-history https://www.eff.org/issues/cda230/legislative-history).

phoehne 4y ago
Thanks. The crux of this is just about what they failed to moderate and if they were acting in good faith?
- jcranmer 4y ago
  The crux is that the plaintiffs want to go after Google in any way they can, and they're grasping at straws to try to eviscerate §230 immunity. (No really, they're arguing that the URLs in search recommendations are original content that therefore doesn't get §230 immunity). §230 is explicitly pretty broad on immunity: it states that providers are neither liable for failing to moderate nor for the consequences of their moderation. "Good faith" is a requirement, but only for actions taken to restrict access to the material (i.e., it doesn't apply to a failure to moderate.)

ajross 4y ago

Could you maybe list some examples less contrived than "we gloriously behead a poultry farmer"? Needless to say completely fictitious partisan warfare is unlikely to convince an actual court. In point of fact there probably is some stuff to worry about, but where we've seen it it's been almost exclusively in off label sites. Unpopular opinion: the core tech portals have been remarkably good stewards of the public good in the face of genuine extremis in the discourse the public wants to engage in. Everyone's got their favorite Moment of Clear Censorship that Must Not Be Permitted, but name any forum and any topic, and I'll be able to show you (with a quick seearch using tools provided by these very tech giants) extensive unimpeded discussion. To wit: SCOTUS isn't going to step in because one tweet got taken down. There are thousands just like it that weren't censored.

phoehne 4y ago
I'm torn on that. I finally just read the actual text instead of relying on my memory of 1996. I get what you're saying, but it isn't a discussion like this one here. It's more like the platforms pick and choose what you're likely to see next. There's not just showing up in the order they were posted or the most relevant things. It might be by how angry they predict it will make me, continue to engage, and get more impressions. You see it in crowds, for example, when people go from having one energetic but good time to being put in a state where they are likely to commit violence. I don't know if Travis Scott is more legend than phenomena, or if the Woodstock '99 people realized they created a metal festival. Human beings are weird creatures and we can do horrific things, given the right context and environment. I think the actual text is rarely a cause for concern, but the manipulation around it is. You can take the same group of people and have an orderly discussion on defenestration, but work them up and you can get an actual defenestration. So if a company is manipulating the flow of information to create an emotional state, on purpose, I'm not sure that company should be 100% in the clear on it. Especially since the law seems to be more about good faith efforts at restricting access to content.