I knew it. I've been saying this is going to become an issue for awhile eg [1][2].
For context Section 230 avoids strict liability for an "interactive computer service" hosting third-party content. For comparison, a first-party site (eg a media site) does have strict liability.
Here's the problem and what (IMHO) will become a big issue: if you, as a platform, only push out content with a particular view, how is that any different to just publishing that view and thus being liable?
Put another way: modern tech companies want to have their cake and eat it too. They want the protections from strict liability and, while hiding behind that shield of third-party content, they analyze the content of posts and they push or suppress posts based on the content. How is that any different to being a first-party publisher?
Example: if 10,000 people write posts for or against South African apartheid and I suppress only those posts that argue against it, aren't I, in effect, publishing a pro-apartheid position?
I think this is a landmine waiting to go off. Tech companies have successfully sold us on the idea that "the algorithm" is somehow agnostic or neutral. Nothing could be further from the truth, particularly with modern AI tools that can deeply classify posts based on their content, politics, etc. Meta's policy chief bragged about doing this [3].
People on HN seem to have this weird, third-rail stance on Section 230. It's nuanced.
Legally, I think making algorithmic decisions based on content should be treated as a first-party act of publishing. And I believe courts are going to get there.
> Legally, I think making algorithmic decisions based on content should be treated as a first-party act of publishing.
That was basically the direction the ruling in Anderson v. TikTok was pointing, yes?
The question is going to be where you draw the line. For example, if HN's automated mechanisms make a post invisible (to anyone who doesn't have showdead enabled) after enough downvotes, is that "making algorithmic decisions based on content"? I don't think so, or at any rate I don't think it's at all the same as the kind of things the big tech giants do algorithmically. But the line needs to be carefully drawn.
>I don't think so, or at any rate I don't think it's at all the same as the kind of things the big tech giants do algorithmically.
It's an algorithmic decision based on content designed to influence the end user, it's exactly the same thing. Section 230 is scale invariant in that it applies to all sites hosting third party content, there is no reason to assume any regulation which replaces it would be any different.
Abolishing 230 means no free speech, anywhere. Speech would need to be restrained, vetted, and approved by a platform, before it could be shared.
While much abusive speech can be blocked from being published with this sort of Prior Restraint, much more healthy, vital speech — that counters abuse — would be repressed by it, IMHO. See how litigation involving copyright law has been used by the Church of Scientology to silence critics, for example. The removal of 230 means platforms themselves could be targeted to efficiently censor large sections of the population.
And the basic principle of Prior Restraint flies in the face of the presumption of innocence that underlies conduct in a Free Society.
That's not true. It would stifle centralized platforms but it would not meaningfully impact those self hosting. So ActivityPub (ie self hosted Mastodon and similar) would be unaffected as would Bluesky with a self hosted PDS (IIUC). Further, paid hosting services for those would be unaffected for the same reason that we don't require section 230 to protect the car rental company if someone uses the vehicle to commit a crime.
I'm not advocating to repeal section 230 but honestly all I'm seeing here is upsides.
> That's not true. It would stifle centralized platforms but it would not meaningfully impact those self hosting.
IDK why you think this. 230 is brief; it doesn't distinguish between size of internet entities. It protects every site that hosts a comment section from bad-faith lawsuits by deep pockets who don't like that speech.
That's all it does. It doesn't provide cover for illegal activity or any of the other 1k things it gets accused of.
>So ActivityPub (ie self hosted Mastodon and similar) would be unaffected as would Bluesky with a self hosted PDS (IIUC).
Only as long as you don't allow third party comments or run a multi-user instance. Then you're just as legally liable for that content as the centralized platforms.
>Further, paid hosting services for those would be unaffected for the same reason that we don't require section 230 to protect the car rental company if someone uses the vehicle to commit a crime.
Mixed metaphors aside, Section 230 is the law that "protects the car rental company if someone uses the vehicle to commit a crime."
If the car rental company is the platform, then prior to Section 230 (and after its repeal) they were and would be held responsible.
Section 230 just establishes that the party posting content is liable for that content, just as the person using a vehicle to commit a crime is responsible for that crime.
It is, in the sense that companies like Meta operate assuming that 230 gives them carte blanche to act as they please.
But:
"Plaintiffs characterize Meta’s duty as one of product design—that Meta should not have built Facebook in a way that boosted incitements to violence. Still, the alleged defects relate to Facebook’s core design as a publishing platform, particularly how Facebook promoted or downplayed third-party posts using algorithms. Under our case law, matching users with content is publishing conduct, even when the user has not requested the content."
At issue the question of who, if anyone, is responsible for the speech issues at play when Meta chooses to show me X rather than Y of the near infinite pool of content at its disposal.
Some are indeed looking to completely get rid of 230, but if you ask me, assuming that Meta has 230 protections for all editorial decisions it makes should be considered overbroad. I'm no lawyer, but as a citizen it seems to me that it puts Meta's influence beyond legal accountability in a way that no individual is. And that seems totally unfair.
This would make every site hosting user generated content vulnerable to expensive lawsuits by authoritarians, deep-pocket censorists and other bad faith actors.
230 enabled most of the internet - specifically the rise of small sites.
Getting rid of 230 would be a massive gift to large platforms who'd be the only ones who could afford the legal bills. Which is why many large platforms have supported ending 230.
So you think that, for example, ycombinator, who provides this forum, should have to be prepared to defend in court every single action it takes that restricts someone's use of this site? (For example, every time a post gets restricted because of downvotes?) Because that's what repealing Section 230 would do.
No, I think that ycombinator, who provides this forum... shouldn't.
This thing where we've traded societal trust for profits for the companies destroying it is a bad choice, and section 230 is what enables it. Let it all burn. Bring back small communities with internal trust relationships. Growth is not an automatic good.
(No, pointing out that I wouldn't be allowed to post this isn't a clever gotcha. The whole point of advocating for change is that it would affect me too.)
> No, I think that ycombinator, who provides this forum... shouldn't.
Good, then we're in agreement on that.
> This thing where we've traded societal trust for profits for the companies destroying it is a bad choice
We're in agreement on that too.
> section 230 is what enables it.
Here is where we disagree. Section 230 is not what enables the big tech giants. Section 230 is what protects sites like this one, that do not do the sorts of things the big tech giants do, but do host user-provided content. Without Section 230, ycombinator would have to be prepared to defend every moderation action it takes in court. The big tech giants actually don't care whether or not they have to do that; the cost of that is hardly even rounding error in their accounting. If Section 230 were repealed, they'd be fine; maybe they'd have to charge some more for ads on average to make up the difference. Big deal.
Sites like this one, however, would not be fine. Ycombinator might still be able to keep this site up, because they're well enough known that they might be able to get some political traction if someone took them to court. But the "small communites with internal trust relationships" that you talk about? Those need something like Section 230 to avoid being sued out of business by the first a*hole that wants to post there and is told they can't because of their behavior, and has enough time and effort to spare to take them to court.
If we want to outlaw something that might put a spoke in the wheels of the big tech giants, IMO it would be the ad-supported business model--the idea that a large corporation providing a service that's of obvious value to billions of people can get out of any responsibility towards those people simply by saying, well, we're providing the service for free, take it or leave it. And then saying they need to monetize those same people's data to stay in business, because, well, how else are we going to pay for this valuable service we're providing?
There are exemptions to section 230 immunity and they have resulted in wild censorship in an effort to avoid liability
So although I functionally agree that responsibility should be shared more, I haven’t seen a framework that accomplishes that. Organizations just stop allowing user generated content entirely, not addressing how they syndicate it
Platforms should bear some responsibility for the content they distribute. Particularly when they run algorithms to promote content based on what they think will engage your attention. There should be no "safe haven."
Still wild to me that people advocate for making platforms responsible for user-submitted content on a website that consists entirely of user-submitted content.
Everyone always has some excuse for why they believe their favorite websites and platforms wouldn’t be impacted, but generally these problems would result in a situation where only the big platforms like Facebook could afford to comply and the platforms like Hacker News or all of your favorite forums and chat hangouts would be forced to close because they can’t afford to comply with laws making them liable for user generated content.
Repealing 230 would only consolidate power among the big social sites who can afford to comply and lobby.
I think the idea that sites have no liability for customly built content designed to capture your attention is kind of wild. It would be one thing if it was a basic or public thumbs up popularity thing based on the groups you join or were apart of. Or a time chronological ordering of things. But its more than that. Its a big math equation that looks at everything you have ever done sells that data and than goes and uses it also to serve you custom ads and contents that maximally steal and maintain attention.
Following you final state and some of the other sub threads, I think removing section 230 hurts larger sites more than smaller ones but also I think its a silly point. If the opposite was true that it benefits small sites more why does 90% of internet traffic go yo just the top 1000 sites? Its because there is no liability and cost to operate in ways that hurt people and the world. A reduced order of the algorithms say chronological ordering or basic vote systems perhaps deserve section 230 protection but the friction roster of videos custom order to capture attention is a type of content production and perhaps needs regulations
Without section 230 you'd get a few large publishing outlets moderating all content and you'd also get distributed media where the aggregation is done by each individual.
There just wouldn't be that middle area where social media sites get to choose what you are allowed to say or read just because they have a monopoly on the data.
For instance, you could run your own individual Hacker News site that collects the data and creates the same thing you see today - except you could choose to only view posts and comments by sources verified as humans. Or you could turn on 'show dead' on a grander scale - your choice.
> For instance, you could run your own individual Hacker News site
It’s baffling how many people think repealing Section 230 would make it easier for small people to host user-submitted content.
It would do the inverse: It would make hosting user submitted content legally untenable for all but the biggest players with the biggest ad budgets to comply with the laws.
You host your own content, and non-230 telecom rules protect a pure cache so bandwidth and always-on internet needn't be an issue.
It shouldn't be baffling to you; it's not even a hard problem to solve, and the only reason why things like ActivityPub aren't done more today is because of 230. A monopoly on data gives centralized social media an unfair advantage that's only possible through legal immunity.
> You host your own content, and non-230 telecom rules protect a pure cache so bandwidth and always-on internet needn't be an issue.
You can already host your own content. Bandwidth is a non-issue right now.
You’re ignoring, or don’t understand, that the benefit of social media is sharing and distribution. The repeal-230 people always have some fanciful ideas about an alternative system appearing to fill the void and that it will be better for reasons, but the bottom line is that without some protection for services to share content it’s going to be negligible exposure.
Only the few with access and connections to the well-connected large media outlets would have distribution. It would be a total self-own for small people who want to share something.
The sites have user submitted content but the sites decide what content they show.
In legacy media the platform is responsible for the content they present you. On new media platforms they are pushing out foreign propaganda and getting financially rewarded for it.
Repeal 230 and put all the controls of recommender system into the hands of the user. Present information chronologically by default.
> Still wild to me that people advocate for making platforms responsible for user-submitted content on a website that consists entirely of user-submitted content.
In both Meta and Google's case, it has been found that it doesn't "consist entirely of user-submitted content", there is also "their software" promoting and recommending content and allowing to pay for higher visibility, and they should certainly be liable for at least this facet of what they do (and were both recently found to have intentionally designed it to be "harmful").
There is also the open question of whether the big tech companies are doing enough to police their platforms, which are all awash with scams, counterfeits and more. Their safeguards for "user-generated content" (somehow including physical goods on Amazon and apps for Android and iPhone) are another facet they should be liable for if it is insufficient.
tl;dr I view Facebook's practices as a form of editorializing but not HN's
Setting aside the valid point about self hosting made by another commenter, I'll note that we don't have to throw the baby out with the bathwater here. Section 230 need not apply to outlets that make "editorial" decisions and can remain in place for places that operate in a largely neutral and transparent manner (for example HN as I understand it).
I see no reason not to hold a tabloid responsible for what it publishes and I further see no inherent reason that highly customized feeds utilizing black box recommendation algorithms designed to benefit the operating party shouldn't be viewed as a form of hands on content promotion (ie editorial in nature).
To me the current arrangement seems like laundering responsibility with "well the algorithm did that, not me" which (IMO) is about as valid as "we didn't violate any financial regulations because we did it all in crypto".
Oh I think sites would be greatly impacted. To the point where most/all user submitted content would need to be moderated. It would completely upend social media and make most comment-based forums impractical to operate at large scale. I think that would be a good thing. We'd be going back to something like the days of newspapers, where letters to the editor were all read and approved by the publisher before they were printed.
You want to go back to an era where only a select few large, powerful outlets could decide what gets published? You see what’s happening with news media and governments threatening news outlets and you want to consolidate to that?
The rest of us would have to beg for their permission to let our words be heard?
I don’t think you’ve thought this fantasy through.
It's good intentions. Only instead of paving the way to hell, it will pave the way to techo fascist governments. Maybe that is hell. Depends on your income bracket.
> Platforms should bear some responsibility for the content they distribute.
My read of section 230 is that they do, and part of that responsibility is remediating bad content as soon as possible?
Whether that is currently being done to the fullest extent is debatable, but I dont see how getting sued into the ground for hosting user content at all even if they remediate it is better?
> and part of that responsibility is remediating bad content as soon as possible?
No. Section 230 protects websites from the liability that they might otherwise bear upon removing objectionable content. You may ask, why would removing objectionable content give someone liability? Because for legal "simplicity" some people want to treat the act of removing harmful content as legal knowledge that the harmful content existed, and if the website was "too slow" to remove harmful content then surely the website should bear responsibility for the harm caused by being "too slow" or possibly complicit.
If there is a non-transparent or non-user changable algorithm selecting content to present then 230 protections shouldn't apply. This is different than removing 230 entirely
Section 230 itself already does not apply to federal criminal investigations, including investigations of child sexual abuse. A social media company should not bear legal responsiblity to detect crimes before law enforcement does.
Counterpoint - Businesses are generally liable in the physical world for harms caused on their premises by a foreseeable hazard. Whether as simple as a pothole in a parking lot that causes injury or an unstable shelf that collapses, …
The major social media sites know absolutely that their products expose customers and users to predictable harm and cause real damages. They should bear an analogous liability burden as a business operating a physical location.
For context Section 230 avoids strict liability for an "interactive computer service" hosting third-party content. For comparison, a first-party site (eg a media site) does have strict liability.
Here's the problem and what (IMHO) will become a big issue: if you, as a platform, only push out content with a particular view, how is that any different to just publishing that view and thus being liable?
Put another way: modern tech companies want to have their cake and eat it too. They want the protections from strict liability and, while hiding behind that shield of third-party content, they analyze the content of posts and they push or suppress posts based on the content. How is that any different to being a first-party publisher?
Example: if 10,000 people write posts for or against South African apartheid and I suppress only those posts that argue against it, aren't I, in effect, publishing a pro-apartheid position?
I think this is a landmine waiting to go off. Tech companies have successfully sold us on the idea that "the algorithm" is somehow agnostic or neutral. Nothing could be further from the truth, particularly with modern AI tools that can deeply classify posts based on their content, politics, etc. Meta's policy chief bragged about doing this [3].
People on HN seem to have this weird, third-rail stance on Section 230. It's nuanced.
Legally, I think making algorithmic decisions based on content should be treated as a first-party act of publishing. And I believe courts are going to get there.
[1]: https://news.ycombinator.com/item?id=47066527
[2]: https://news.ycombinator.com/item?id=46148529
[3]: https://theintercept.com/2024/10/21/instagram-israel-palesti...
That was basically the direction the ruling in Anderson v. TikTok was pointing, yes?
The question is going to be where you draw the line. For example, if HN's automated mechanisms make a post invisible (to anyone who doesn't have showdead enabled) after enough downvotes, is that "making algorithmic decisions based on content"? I don't think so, or at any rate I don't think it's at all the same as the kind of things the big tech giants do algorithmically. But the line needs to be carefully drawn.
It's an algorithmic decision based on content designed to influence the end user, it's exactly the same thing. Section 230 is scale invariant in that it applies to all sites hosting third party content, there is no reason to assume any regulation which replaces it would be any different.
While much abusive speech can be blocked from being published with this sort of Prior Restraint, much more healthy, vital speech — that counters abuse — would be repressed by it, IMHO. See how litigation involving copyright law has been used by the Church of Scientology to silence critics, for example. The removal of 230 means platforms themselves could be targeted to efficiently censor large sections of the population.
And the basic principle of Prior Restraint flies in the face of the presumption of innocence that underlies conduct in a Free Society.
I'm not advocating to repeal section 230 but honestly all I'm seeing here is upsides.
IDK why you think this. 230 is brief; it doesn't distinguish between size of internet entities. It protects every site that hosts a comment section from bad-faith lawsuits by deep pockets who don't like that speech.
That's all it does. It doesn't provide cover for illegal activity or any of the other 1k things it gets accused of.
Only as long as you don't allow third party comments or run a multi-user instance. Then you're just as legally liable for that content as the centralized platforms.
>Further, paid hosting services for those would be unaffected for the same reason that we don't require section 230 to protect the car rental company if someone uses the vehicle to commit a crime.
Mixed metaphors aside, Section 230 is the law that "protects the car rental company if someone uses the vehicle to commit a crime."
If the car rental company is the platform, then prior to Section 230 (and after its repeal) they were and would be held responsible.
Section 230 just establishes that the party posting content is liable for that content, just as the person using a vehicle to commit a crime is responsible for that crime.
I think Mike Masnick has a superb grasp of the salient issues, and is articulate. I recommend his writing on the topic.
https://www.techdirt.com/2021/10/29/everything-you-know-abou...
But:
"Plaintiffs characterize Meta’s duty as one of product design—that Meta should not have built Facebook in a way that boosted incitements to violence. Still, the alleged defects relate to Facebook’s core design as a publishing platform, particularly how Facebook promoted or downplayed third-party posts using algorithms. Under our case law, matching users with content is publishing conduct, even when the user has not requested the content."
At issue the question of who, if anyone, is responsible for the speech issues at play when Meta chooses to show me X rather than Y of the near infinite pool of content at its disposal.
Some are indeed looking to completely get rid of 230, but if you ask me, assuming that Meta has 230 protections for all editorial decisions it makes should be considered overbroad. I'm no lawyer, but as a citizen it seems to me that it puts Meta's influence beyond legal accountability in a way that no individual is. And that seems totally unfair.
This would make every site hosting user generated content vulnerable to expensive lawsuits by authoritarians, deep-pocket censorists and other bad faith actors.
230 enabled most of the internet - specifically the rise of small sites.
Getting rid of 230 would be a massive gift to large platforms who'd be the only ones who could afford the legal bills. Which is why many large platforms have supported ending 230.
This thing where we've traded societal trust for profits for the companies destroying it is a bad choice, and section 230 is what enables it. Let it all burn. Bring back small communities with internal trust relationships. Growth is not an automatic good.
(No, pointing out that I wouldn't be allowed to post this isn't a clever gotcha. The whole point of advocating for change is that it would affect me too.)
Good, then we're in agreement on that.
> This thing where we've traded societal trust for profits for the companies destroying it is a bad choice
We're in agreement on that too.
> section 230 is what enables it.
Here is where we disagree. Section 230 is not what enables the big tech giants. Section 230 is what protects sites like this one, that do not do the sorts of things the big tech giants do, but do host user-provided content. Without Section 230, ycombinator would have to be prepared to defend every moderation action it takes in court. The big tech giants actually don't care whether or not they have to do that; the cost of that is hardly even rounding error in their accounting. If Section 230 were repealed, they'd be fine; maybe they'd have to charge some more for ads on average to make up the difference. Big deal.
Sites like this one, however, would not be fine. Ycombinator might still be able to keep this site up, because they're well enough known that they might be able to get some political traction if someone took them to court. But the "small communites with internal trust relationships" that you talk about? Those need something like Section 230 to avoid being sued out of business by the first a*hole that wants to post there and is told they can't because of their behavior, and has enough time and effort to spare to take them to court.
If we want to outlaw something that might put a spoke in the wheels of the big tech giants, IMO it would be the ad-supported business model--the idea that a large corporation providing a service that's of obvious value to billions of people can get out of any responsibility towards those people simply by saying, well, we're providing the service for free, take it or leave it. And then saying they need to monetize those same people's data to stay in business, because, well, how else are we going to pay for this valuable service we're providing?
So although I functionally agree that responsibility should be shared more, I haven’t seen a framework that accomplishes that. Organizations just stop allowing user generated content entirely, not addressing how they syndicate it
Everyone always has some excuse for why they believe their favorite websites and platforms wouldn’t be impacted, but generally these problems would result in a situation where only the big platforms like Facebook could afford to comply and the platforms like Hacker News or all of your favorite forums and chat hangouts would be forced to close because they can’t afford to comply with laws making them liable for user generated content.
Repealing 230 would only consolidate power among the big social sites who can afford to comply and lobby.
Following you final state and some of the other sub threads, I think removing section 230 hurts larger sites more than smaller ones but also I think its a silly point. If the opposite was true that it benefits small sites more why does 90% of internet traffic go yo just the top 1000 sites? Its because there is no liability and cost to operate in ways that hurt people and the world. A reduced order of the algorithms say chronological ordering or basic vote systems perhaps deserve section 230 protection but the friction roster of videos custom order to capture attention is a type of content production and perhaps needs regulations
There just wouldn't be that middle area where social media sites get to choose what you are allowed to say or read just because they have a monopoly on the data.
For instance, you could run your own individual Hacker News site that collects the data and creates the same thing you see today - except you could choose to only view posts and comments by sources verified as humans. Or you could turn on 'show dead' on a grander scale - your choice.
Section 230 isn't required for social media.
It’s baffling how many people think repealing Section 230 would make it easier for small people to host user-submitted content.
It would do the inverse: It would make hosting user submitted content legally untenable for all but the biggest players with the biggest ad budgets to comply with the laws.
It shouldn't be baffling to you; it's not even a hard problem to solve, and the only reason why things like ActivityPub aren't done more today is because of 230. A monopoly on data gives centralized social media an unfair advantage that's only possible through legal immunity.
You can already host your own content. Bandwidth is a non-issue right now.
You’re ignoring, or don’t understand, that the benefit of social media is sharing and distribution. The repeal-230 people always have some fanciful ideas about an alternative system appearing to fill the void and that it will be better for reasons, but the bottom line is that without some protection for services to share content it’s going to be negligible exposure.
Only the few with access and connections to the well-connected large media outlets would have distribution. It would be a total self-own for small people who want to share something.
In legacy media the platform is responsible for the content they present you. On new media platforms they are pushing out foreign propaganda and getting financially rewarded for it.
Repeal 230 and put all the controls of recommender system into the hands of the user. Present information chronologically by default.
In both Meta and Google's case, it has been found that it doesn't "consist entirely of user-submitted content", there is also "their software" promoting and recommending content and allowing to pay for higher visibility, and they should certainly be liable for at least this facet of what they do (and were both recently found to have intentionally designed it to be "harmful").
There is also the open question of whether the big tech companies are doing enough to police their platforms, which are all awash with scams, counterfeits and more. Their safeguards for "user-generated content" (somehow including physical goods on Amazon and apps for Android and iPhone) are another facet they should be liable for if it is insufficient.
Setting aside the valid point about self hosting made by another commenter, I'll note that we don't have to throw the baby out with the bathwater here. Section 230 need not apply to outlets that make "editorial" decisions and can remain in place for places that operate in a largely neutral and transparent manner (for example HN as I understand it).
I see no reason not to hold a tabloid responsible for what it publishes and I further see no inherent reason that highly customized feeds utilizing black box recommendation algorithms designed to benefit the operating party shouldn't be viewed as a form of hands on content promotion (ie editorial in nature).
To me the current arrangement seems like laundering responsibility with "well the algorithm did that, not me" which (IMO) is about as valid as "we didn't violate any financial regulations because we did it all in crypto".
The rest of us would have to beg for their permission to let our words be heard?
I don’t think you’ve thought this fantasy through.
Only the large powerful outlets will have the budgets and legal connections to host user submitted content.
My read of section 230 is that they do, and part of that responsibility is remediating bad content as soon as possible?
Whether that is currently being done to the fullest extent is debatable, but I dont see how getting sued into the ground for hosting user content at all even if they remediate it is better?
No. Section 230 protects websites from the liability that they might otherwise bear upon removing objectionable content. You may ask, why would removing objectionable content give someone liability? Because for legal "simplicity" some people want to treat the act of removing harmful content as legal knowledge that the harmful content existed, and if the website was "too slow" to remove harmful content then surely the website should bear responsibility for the harm caused by being "too slow" or possibly complicit.
who knows.
https://time.com/7336204/meta-lawsuit-files-child-safety/
The major social media sites know absolutely that their products expose customers and users to predictable harm and cause real damages. They should bear an analogous liability burden as a business operating a physical location.
Sounds great. Last I checked, most businesses don't have a legal obligation to police the conversations that happen on or around their premises.