[EdJiang]

If you haven't read Section 230, go do so now. It's enabled the development of the modern internet as we know it, and the meat is only 3 sentences. The rest is preamble or interactions with other laws.

> (c) Protection for "Good Samaritan" blocking and screening of offensive material

> (1) Treatment of publisher or speaker

> No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.

> (2) Civil liability

> No provider or user of an interactive computer service shall be held liable on account of-

> (A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected; or

> (B) any action taken to enable or make available to information content providers or others the technical means to restrict access to material described in paragraph (1).

https://uscode.house.gov/view.xhtml?req=(title:47%20section:...

[TimPC]

There is good reason to reform section 230. Right now courts are applying the liability so broadly that companies aren’t liable even after they are notified about illegal behaviours on their site. In a court case involving Grindr refusing to take down a profile created by someone’s ex-bf that was being used to harass him, their refusing to so even after contact by lawyers was protected under section 230 and the case was thrown out. I’d be all for a modified version of section 230 that required sites to have a contact email and made them liable if they don’t address certain issues in an appropriate time period.

It’s also worth mentioning that before section 230 if you didn’t moderate you weren’t liable so in certain senses it’s a censorship bill rather than a free speech one since it protects removing speech. That being said I do understand the need to moderate sites and remove some content, hence my proposal of the modified version rather than a call for its elimination entirely.

[eli]

> a modified version of section 230 that required sites to have a contact email and made them liable if they don’t address certain issues in an appropriate time period.

So recreate the DMCA Takedown process but for speech? Do you think the DMCA is working well for copyright holders and users?

The abuse of this would be massive. Let's say I don't like the comments you wrote so I email the host of the forum they're on and say they're defamatory. Now the host has to decide if they are defamatory (which is often a tough call even for lawyers) and also weigh the risk that I might file a costly lawsuit anyway. Or they just delete the comment.

[hajile]

They always have the option of allowing everything and ignoring emails -- something the dcma doesn't offer.

[s__s]

To me that problem can be dealt with between the user, their ex-bf and local law enforcement. Grindr need not be involved, and no special internet laws need apply.

That’s a case of harassment and possibly some form of identity theft.

I don’t find your proposal tenable and it would obviously be prone to abuse.

[gowld]

That's only true of Grinder is obligated to positively ID a US resident who provided the content.

[odessacubbage]

there are a lot of reasons why grindr would not want to do that. grindr is not tinder and it's not okcupid a BIG part of the service is providing psuedonymity to it's users. attitudes about homosexuality in the us and especially internationally make it a pretty significant liability for a service to start 'outing' people. if you can be ID'ed in the US you can be ID'ed in singapore or chechnya. and even besides that there's a great degree of cultural complexity at play and even in 'tolerant' places many men are discreet because they don't want to be branded as a fag. if i ran grindr i would've made the same call

[dragonwriter]

> Right now courts are applying the liability so broadly that companies aren’t liable even after they are notified about illegal behaviours on their site.

Right. While the text of the bill doesn't remove distributor liability (only publisher/speaker), its applied as doing that to, and giving even actively-moderating sites only neutral platform liability. There may be justification for this in legislative history and legal construction, so it may not be a pure judicial mistake, but from a policy perspective its at least arguably an overcompensation that Congress should correct, a correction which would be much more modest than many of the reform/repeal Section 230 proposals but probably hit a better point in terms of dealing with the worst problems without creating more than it solves.

> It’s also worth mentioning that before section 230 if you didn’t moderate you weren’t liable

That's not entirely true. If you did actively moderate, you were liable as publisher, but if you didn't actively moderate you would still likely be held liable as a distributor.

> in certain senses it’s a censorship bill

The entire Communications Decency Act was a censorship bill and the express purpose of 230 as part of the CDA was to encourage sites to do moderate content instead of taking a hands-off position.

OTOH, so long as there is liability for knowing-unlewful content (distributor liability) and antitrust enforcement, I think that's a good thing and reduces the social pressure for government to push the maximum line the courts will let it get away with in terms of government content restrictions.

[curryst]

Grindr didn't refuse to take them down; the ex-bf kept creating new ones. That's a whole different problem. Grindr claims they were monitoring for new profiles, but that some slipped through their checks.

In that scenario, I don't know what a reasonable level of effort for Grindr to exert is. It seems infinitely unreasonable to make them liable for any failure; there is a determined person on the other end that will probably eventually find some way of adding spaces or using symbols instead of letters, or using weird UTF-8 symbols or something.

I don't see Grindr as failing there; while they probably could have done more, they seem to have made a best faith effort to stop it. The police should have intervened and filed charges against the boyfriend for stalking and harassment. Even failing that, I would have filed a civil case so I could subpoena the logs from Grindr and used them as evidence in a restraining order.

Grindr is not the appropriate party to resolve this. I don't call Ford when people drive their trucks like assholes. I don't call Glock when somebody shoots someone. If you're going to call Grindr, you might as well call their ISP and Google too, see if you can get the ISP to block Grindr or get Google to route Grinder to localhost. They're complicit in enabling this too.

> Right now courts are applying the liability so broadly that companies aren’t liable even after they are notified about illegal behaviours on their site

This, to a degree, makes sense. They haven't been notified about illegal behavior on their site, they have been notified of allegedly illegal behavior on their site. Grindr is well within their rights to say that they don't believe that the profile violates any laws. For example, it says that he attempted to file for a restraining order and was denied. So that court either found that what the ex-bf was doing wasn't illegal, or that he failed to meet the requirement of a preponderance of evidence. So he failed to convince a judge that his ex was more likely than not stalking him. Should Grindr be required to take action on a claim that is more likely false than true?

> I’d be all for a modified version of section 230 that required sites to have a contact email and made them liable if they don’t address certain issues in an appropriate time period.

That's fraught with issues. What counts as addressing the issue? Is it banning the profiles as people identify them? Is it banning the personal info from appearing in profiles? Do they have to hire a group of people to memorize all the bits of bad data, and check new profiles and profile updates for those snippets, as well as any clever encodings that a computer wouldn't recognize?

What is an appropriate time period? Is it some flat period, like a week, regardless of what changes are required? Does it vary, and if so, who decides what's a reasonable amount of time?

This is not to mention that literally none of this goes through a court, which is terrifying and exceptionally prone to abuse. Of course, it could go through a court, but we already have laws and remedies for this situation in court.

Cases like that make it seem really cut and dry, like there would never be a grey area. Even ignoring cases of outright fraud, what do you do in situations where one side feels victimized but it doesn't actually meet any legal standards? Like if person A always replies and argues with person Bs tweets. When person B blocks person A, they make a new account. Person B says they feel harassed and wants to force Twitter to do something about it. Person A says that Twitter is a public forum, and that if people don't want other people to disagree, they should use a more private forum. It never goes further than that. No threats, no doxxing, no real life interactions. Person A is probably an asshole, sure, but I don't think section 230 grants you immunity from assholes. I don't think it counts as stalking or harassment either (though I could certainly be wrong, not a lawyer). Should we really allow Person B to force Twitter to do something without having a judge involved? I would really rather not give the Twitter lynchmobs yet another way to dispense their own vigilante justice.

[odessacubbage]

not only that, plenty of people just use blank accounts with zero info and only exchange pics and identifying information in dms. the malicious ex could just do that with the same effect. if i'm a psycho bitch who scrawls your name/number and compromising information in hundreds of truckstop bathrooms across the tristate, are shell and conoco liable? that would be absurd and it's equally as absurd in this case.

[wahern]

I don't think it would be world ending to get rid of Section 230. I almost would like to see it happen, if only because it would have precisely the opposite effect expected by all the people whining about being censored. Though, I suppose you can't be censored if the channel itself is extinguished.

More practically, the technically literate would go back to the world of Usenet, mailing-lists, and minimalistic forums like HN, hopefully inventing distributed reputation systems in the process. I have this vague idea for PGP web of trust-like signing of Usenet posts (published as hidden posts when readers +1/-1) which are then SPAM scored based on the depth of the attestation chain to the reader's own trusted posters, which may have been seeded from one or more centralized databases of group maintainers, similar to the current registration system for moderated Usenet groups except you could freely choose alternative registrars.

[NationalPark]

Would forums like HN survive? I can think of a few incidents where malicious information about people made the front page then turned out to be false. Is HN prepared to defend against lawsuits about that? Is HN prepared to lose lawsuits about that?

It sounds like you're basically suggesting that making the internet useless is a good thing, because maybe something something cool will come out of the ashes and there's a chance it could be even better after a bunch of extremely hard and broad problems are solved. I don't like those odds.

[wahern]

HN already has moderators who do a very good job of filtering posts in a timely manner. HN's exposure to liability for libel would be rather minimal. People and companies are exposed to legal risk all the time, everywhere they go, and somehow they don't curl up into a ball and die of starvation in their basements.

Big, diverse sites like Facebook and Twitter need Section 230 because they can't effectively use human moderators to sift through the content. They have to rely on machine learning, which has false negative rates magnitudes higher than a human. Yet at the same time, they're constantly trying to shape and edit and, basically, narrate the user content, as part of their monetization strategy. That's their dilemma.

Moreover, the distinction between publisher and distributor will still exist. The alternative to strong moderation is no moderation--you're just a distributor, like a Usenet node or the telephone company. But that's more difficult to monetize. (Of course, the legal landscape would be more nuanced than that--traditional libel law wouldn't demand a simple dichotomy between moderation and no moderation.)

Without Section 230 companies would have a more difficult time trading profit potential for legal liability, but it would still be done. Newspapers, write-in columns, bulletin boards, and other forums were around for centuries, all the same exposed to libel law. Even the internet was around for decades prior to Section 230.

[eli]

> HN's exposure to liability for libel would be rather minimal.

I don't understand how you reached that conclusion. They could be sued over any comment that appears for any amount of time. There are definitely comments that have appeared on HN that are libelous.

Moreover even if they pre-screened every comment before it was posted with a team of lawyers who never make any mistakes, they'd STILL have to worry about defending against frivolous lawsuits. Would it even be possible to buy liability insurance for a forum in this world? It would cost a fortune.

And this is for a site that has the resources to have full time moderators. Smaller sites are even worse off.

I don't see how anyone could practically operate any forum or discussion board or comments section that allowed people to post messages in real time.

>Even the internet was around for decades prior to Section 230.

Sure and sometimes your ISP got successfully sued because someone didn't like a comment posted on a message board they hosted.

[nullc]

> have appeared on HN that are libelous

Not just have appeared, but which are still on display.

Sometimes the difference between libellous and a critical statement protecting the public is purely the difference of the statement being true or not.

This is not something a moderator is necessarily in a position to be able to judge but it's critically important to a community that its members can communicate true negative facts about other members.

[curryst]

> Sometimes the difference between libellous and a critical statement protecting the public is purely the difference of the statement being true or not.

Or whether the person making the statement knew that it was false at the time (or should have known). That won't protect the statement from being libelous, but it limits your maximum liability to actual damages that you can show (which in many cases is only going to be the lawyer fees; in the case of a widely repeated libelous statements, how do you determine what harm came from which sites?).

> This is not something a moderator is necessarily in a position to be able to judge

I would make that a harder statement. Moderators cannot tell whether a certain post is libelous. Even assuming that the moderator knows whether the post is factually true or not (and there are a lot of accusations where, when they come out, no one knows for sure who is telling the truth), whether it's libelous depends on whether the person in question is considered a public figure, and whether the person that posted it did enough fact checking to be deemed sufficient in attempting to prove or disprove it. The only person that can determine whether the person being defamed is a public figure, and whether the burden to verify the facts was met, is a judge with jurisdiction over the case. Anything else is just people guessing at how a judge would interpret this case, which is fraught with problems, including and up to that two judges who have jurisdiction would disagree on some facet of that.

[eli]

Worse yet: sometimes the difference is just who happens to be on your jury that day!

[wahern]

> They could be sued over any comment that appears for any amount of time. There are definitely comments that have appeared on HN that are libelous.

1) You could be sued now for [potentially] libelous comments you write on HN. What's the average wealth of HN posters? How many times has HN had to field user account disclosure requests so commenters could be sued?

2) There are scenarios where HN could be sued now for [potentially] libelous material. For example, in the way moderators reword titles. Not sure how likely they would be to succeed, but it's certainly plausible, and it would be relatively cheap for a lawyer to test the waters. I'd be curious to see how many letters Y Combinator has had to field regarding its content. I suspect greater than 0, but still relatively few. Do its lawyers toss them in the trash, discounting to $0 the risk of liability? I doubt it--while they may consider the risk low, it's still something, and that something presumably effects HN's policies today.

A few months ago I learned a memorable phrase from an HN comment: think in probabilities, not possibilities. Regarding Section 230, most people seem to be in a mode of thinking where they simply compare a world with existentially oppressive liability vs no liability whatsoever. The world doesn't work that way, not even U.S. law. We're all subject to the possibility of financially existential liability every time we drive a car, but we're not crippled by it. How many Silicon Valley engineers with million-plus dollar homes and assets even have umbrella coverage? While I suspect the number is far fewer than what would be rationally called for, the reason is nonetheless because the probabilities are far less ominous than the possibilities.

Would HN's liability exposure grow? Absolutely. Would their legal costs, including possible settlements, increase? I would think. How would the site change? It's hard to say, but I'll go on record as saying that I don't think it'd be taken down, and I seriously doubt there would be many, if any, substantive changes to current policies and practices.

> Sure and sometimes your ISP got successfully sued because someone didn't like a comment posted on a message board they hosted.

To be clear, my only claim is that I don't think it would be the end of the internet or even social media. It might be the end of Twitter and Facebook as we know it, but the U.S. grants to participatory websites one of the, if not the strongest defenses to libel liability in the developed world, and yet the internet works much the same everywhere else lacking such a strict defense. Likewise, many people consider civil tort liability entrepreneurially oppressive in the U.S., and yet private enterprise--grocery stores, manufacturers, schools, etc--exist much the same here as they do elsewhere, especially in other developed countries. In fact, often they willfully subject themselves to more risk than they would elsewhere. (That's one benefit of a system that relies on private suits as opposed to regulatory mandates or criminal sanctions.) And yet the worst figures I've seen for the supposed comparative cost to the immensely successful U.S. economy of it's overly litigious civil legal system is something like 5% of GDP.

There's alot of hyperbole and hand-wringing surrounding this issue, and a big reason, IMO, for it relates to our contemporary, radical narratives regarding Free Speech on the one hand and American litigiousness on the other. While anxiety regarding both may be rooted in a kernel of truth, the full truth and reality--legal, political, social--doesn't support the extreme reactions and doomsday predictions.

While I'm not advocating for repeal of Section 230, I'd trade it in a heartbeat for legislative voiding of Qualified Immunity, if that sort of compromise was on the table between Democrats and Republicans. That's the sort of flexible, pragmatic thinking I wish there was more of in our public discourse. But it can't happen if we're all single-issue voters on every issue, which is what absolutist, possibility-not-probability thinking has turned us into.

[NationalPark]

We won't know the actual exposure until it ends up in the courts. Remember, we're talking about removing the good faith liability protections. Maybe a few thousand views of a libelous comment is enough, even if it was eventually removed. Either way, someone has to hire lawyers to go defend this, so it's not free.

We should also expect new bad actors to take advantage of this. As long as they can spam libel faster than moderators can delete it, they can force the site to shut down or risk the lawsuits. While I'm sure YC has its share of enemies deserved or not, even perfectly innocent people are attacked online every day for no reason at all.

[gowld]

There's no reason to assume that an operator would be liable for libel spam, since they lack mens rea intent. Libel would only be in play if they intentionally refused to take down content or tried to extort people with it.

[ColanR]

I think the odds are pretty good. There's a lot of smart & motivated people who really like the internet, who would probably go a long way to replace it.

[gowld]

Why aren't those people interested in working on that today?

[throwawaygh]

The plan is not to repeal Section 230. The plan is to make protection contingent on appeasing political appointees at the FTC.

Whoever controls the FTC will be able to (and will) pressure the major social media networks into acting as a propaganda arm for their political party.

As dystopian as FB and Twitter are today, in this case, the medicine is poison.

See https://www.hawley.senate.gov/senator-hawley-introduces-legi...

[eli]

Hawley's plan is just one of them.

Some people, including both Joe Biden and Donald Trump, have called for a complete repeal of Section 230 at various times in the last year.

[dragonwriter]

> Some people, including both Joe Biden and Donald Trump, have called for a complete repeal of Section 230 at various times in the last year

AFAICT, that characterization of Biden's position is based entirely on a single oral interview response, which quite arguably was not saying that the law should be repeated but that, on the facts of Facebook's specific conduct, and that of some unspecified other platforms, their conduct should be excluded from Section 230 protections because they were knowingly engaging in misinformation.

Note that Section 230 protections in case law are broader than what is provided on the face of the statute; in addition to the "publisher or speaker" protection in Section 230(c)(1); courts have extended it to also prevent liability as a distributor for content, IIRC by synthesizing 230(c)(1) and the good-faith blocking rule in Section 230(c)(2) and some legislative history to add the not-express-in-statute rule that sites are also not liable even as a distributor for the material they don't block, with some exceptions. Biden's statement is consist with restricting Section 230 to what is says on the face, which would be only removing publisher/speaker liability, not distributor liability (which comes about when the distributor has knowledge or legal notice of the legal problem with the content.)

[wbl]

IANAL but The First Amdment, not Section 230, is what lets me say that Donald Trump is in league with reptilians to enslave all Americans who prefer pork to beef.

Facebook spreading the above, or other similarly ludicrous information, is likewise protected.

[dragonwriter]

> IANAL but The First Amdment, not Section 230, is what lets me say that Donald Trump is in league with reptilians to enslave all Americans who prefer pork to beef.

The First Amendment does not protect you saying that if it is false and you have knowledge that it is false or are grossly reckless in saying it without confirming its veracity, see, New York Times v. Sullivan.

Section 230 is what prevents Facebook from sharing your liability, as a publisher, if they relay your saying that in the conditions in which you would be liable for defamation.

[eli]

I think that's a very generous reading of what Biden said, especially considering the followup question and the fact that he's declined to clarify his position in the intervening 8 months. Search "230" on this page to see it https://www.nytimes.com/interactive/2020/01/17/opinion/joe-b...

Anyway my point was that there have been calls to repeal 230 from across the ideological spectrum.

[dragonwriter]

> I think that's a very generous reading of what Biden said, especially considering the followup question

In the followup, Biden reiterates the conduct condition and the knowing falsehood criteria, which reinforces rather than weakens the impression that he is calling for the protections of Section 230 to be inapplicable to the actor/action in question due to their knowledge, a distributor-like standard, and not for the law itself to be repealed generally.

I suppose you could read the first line of his response to the second followup ("He should be submitted to civil liability and his company to civil liability, just like you would be here at The New York Times") as calling for publisher-like liability if you ignore the explicit references to actual knowledge as the basis for nonprotection in both the original response and the first followup, but I do think that that is the more strained interpretation, not the less strained.

> and the fact that he's declined to clarify his position in the intervening 8 months.

Why would you assume that he doesn't want to clarify because he wants a full repeal? Its not as if there isn't a constituency for a full repeal, especially on the right, and a key part of Biden's strategy is holding together a Bernie Sanders-to-Bill Kristol left-right alliance against Trump. Keeping disagreements the details of his position on the issue (which is clearly peripheral to his platform, on the grand scheme of things) out of the reasons for people to not feel comfortable with him is as plausible a motivation for that regardless of which side of the full-repeal-vs.-reform his preference on 230 sits on.

[throwawaygh]

Do you have links to actual plans from other folks? Hawley's is the only one I can find an actual draft bill for.

[eli]

Brian Schatz and John Thune have one: https://www.schatz.senate.gov/imo/media/doc/OLL20612.pdf

The DOJ has one: https://www.justice.gov/opa/pr/justice-department-unveils-pr...

The Whitehouse has a somewhat bogus EO https://www.whitehouse.gov/presidential-actions/executive-or...

There have been a bunch of attempts to rewrite and at least a few attempts to just repeal it from both Democrats and Republicans.

[throwawaygh]

Thanks.

[adamiscool8]

That's nonsense. Showing "their algorithms and content-removal practices are politically neutral" is not an insurmountable bar. It's just inconvenient for Big Tech's supporting interests.

[throwawaygh]

Really? You think we can here in this thread all agree to what it means for an algorithm or content-removal practice to be "politically neutral"?

If so, please go ahead! But I seriously doubt it. This is a thing political philosophers argue about in journals to this day, that lawyers argue about in SCOTUS cases to this day, and that has been litigated to death in thousands of HN threads over the years.

The question of what "politically neutral" means is perhaps the MOST political question there is. The delineation of political speech from non-political speech defines the playing field.

And even setting aside genuine disagreement, politics does not operate on good faith. It operates on power. In practice, the bill does not outline specific criteria. So "politically neutral" will mean whatever the FTC wants it to mean. Which means it will mean whatever the appointees of the FTC chair want it to mean.

Josh Hawley, of course, knows and understands how power works. He would not be proposing this bill if the big tech companies were right-biased. Democrats also understand how power works. So, in this counter-factual world of right-biased social media, it would be Democrats clamoring for federal intervention and Hawley decrying the "Democrat attack on the most successful American companies". Do you really believe otherwise?

[JoshTriplett]

Exactly. Also, people have a stronger uncomfortable negative reaction to news they don't like than a positive reaction to news they agree with, and extremists think even neutral descriptions of reality are biased against them, so if you show a neutral selection to a non-neutral person they're likely to see it as biased because it doesn't align with their perception of what the proportions should be. Nobody will ever agree on what "neutral" means, which means most likely it'll mean "biased in favor of who currently has political power".

[judge2020]

Exactly. Imagine a republican-backed FCC arguing that, because more voters in the US are democrat, it's not politically neutral to try to show a news article to everyone in the US and the algorithm should try to show it only to an equal number of republicans and democrats.

[dragonwriter]

> You think we can here in this thread all agree to what it means for an algorithm or content-removal practice to be "politically neutral"?

Well, there's a simple answer, but I doubt we'll agree on it. It is impossible for a content-removal practice, algorithmic or otherwise, to be politically neutral. Any such practice will involve (whether implemented case-by-case or encoded into the design of the algorthm) judgements of a political nature and with political impacts.

[throwawaygh]

> It is impossible for a content-removal practice, algorithmic or otherwise, to be politically neutral.

Right. My point is that Hawley's whole premise of a "politically appointed political neutrality committee" is absurdly transparent.

[whimsicalism]

I'm not a republican, nor a Trump supporter, but I disagree. I think the courts would be able to create a body of case law over whether a removal was due to a post being "violent, obscene or harassing" or for some other reason.

We can't be terrified of regulating platforms that have massive amounts of control over what most people see or hear about.

[throwawaygh]

> I think the courts would be able to create a body of case law over whether a removal was due to a post being "violent, obscene or harassing" or for some other reason.

1. Maybe, but that's not what Hawley's bill does.

2. Leaving inherently political questions up to the courts invites politicizing the courts -- something that's already happened and that, if it continues apace, threatens to delegitimize and gridlock the entire federal legal system.

3. Given that you're not a Trump supporter or Republican, perhaps you should review the last 20 years of federal judicial appointments before placing so much faith in the courts...

> We can't be terrified of regulating platforms that have massive amounts of control over what most people see or hear about.

Agreed. I think there are lots of reasonable approaches toward regulation and/or self-regulation. The ability of customers to choose from a marketplace of recommendation algos (or implement their own) is the obvious market-based solution.

However, I do not think a politically appointed committee whose job is to define political neutrality is a reasonable approach. And I think that leaving inherently political moderation choices up to the courts would be even worse -- at least FTC chairs aren't lifetime appointments, and at least politicizing the FTC won't deteriorate public trust in the one portion of the federal government that is not yet perceived as nakedly partisan.

[adamiscool8]

Very simple. No primary moderation action should be made based on human input. Automated moderation should look for identifiable harms (i.e. illicit content, directed threats, terrorism), and absolutely nothing should be removed or blocked based on vague and nebulously defined concerns over "misinformation". Voila -- political neutrality in moderation.

[throwawaygh]

> No primary moderation action should be made based on human input

1. That means no HN.

2. I normally don't have to remind people of this at places like HN, but... algorithms are written by... humans! Supervised algos use data labeled by... humans!

> Automated moderation should look for identifiable harms (i.e. illicit content, directed threats, terrorism)

Why do you list terrorism separately from directed threats?

What is the line/difference between "terrorism" and an "undirected threat"?

Are militia groups that don't make directed threats terrorists? Are radical religious groups that don't make directed threats terrorists? What if they are run by actual terrorists but none of the speech amounts to a directed threat?

Speaking of which, what is a terrorist organization? Is the KKK? What about small white nationalist or black power militia groups? What about QAnon? What about antifa? What about BLM? What about Westboro Baptist? What about the Black Panthers?

There are people -- elected officials -- who think each of those is a terror organization.

So, defining terrorist organization is absolutely a political fight. Maybe we avoid that and just talk about directed threats/ Ok. Does that mean that Al Qaeda allowed to operate on FB as long as they don't make directed threats? In fact, that FB is prohibited from not allowing Al Qaeda on as long as they don't make directed threats? That seems like not a solution anyone is going to get behind.

We haven't even gotten past the "obviously terrorism=bad" and we already have to declare whether BLM, QAnon, Westboro, or militia groups are "terrorists". Which some senators believe is the case and is a 100% political question.

> illicit content

Is Ginsberg's Howl illicit? Is a picture of two women kissing illicit? What about non-sexualized nude breasts? What about nude male bodies? What about an erect penis but in a non-erotic context? Will the dominant answers to these questions be the same in 50 years?

Lots of people would say a site that allows pictures of heterosexual kissing but not not pictures of homosexual kissing is obviously taking a political position, but that was outside the realm of "political opinion" when I entered adulthood! Any public homosexual display of affection was obviously illicit.

> absolutely nothing should be removed or blocked based on vague and nebulously defined concerns over "misinformation".

What does vague mean? What does nebulously defined mean? What is the difference between misinformation and libel? What is the difference between misinformation and dangerous information? Is it impressible to remove a video that's targeted at kids and encourages huffing glue as a fun and harm-free activity?

Anyone who has moderated a forum knows that such an algorithm is going to have all sorts of holes and perceived biases. I've never written an automod that some user doesn't get pissed off about.

More generally: that's just straight-up moderation, it has nothing to do with tweaks to recommendation algos.

What if Twitter realizes that people leave the site if they see stuff about abortion but stay if they see stuff about LGBT rights? Again, viewpoint-neutral, Americans just one day start yawning about abortion and really polarize on LGBT stuff. Can they prioritize posts about LGBT rights over posts about abortion as long as the content served up on the preferred topic is viewpoint-neutral and the only algorithmic goal is more lingering eyeballs?

If no to that, how about sports news vs. SCOTUS decision news?

If yes to that, what about COVID case counts vs. Jobs Report numbers?

Even more generally: anyone who's stayed up to date on robust machine learning knows that defining good notions of robustness -- and political neutrality is a type of robustness -- is very much an open problem. So even if we had a precise definition of political neutrality, which I don't think we do, "simply create an algorithm that has that property" is very much an open algorithmic problem.

In fact, there are even some impossibility theorems in this space. So even if we can define neutrality in a perfectly neutral way -- which we can't -- this might be like passing a constitutional amendment that demands a voting system has all of: Non-dictatorship, unrestricted domain, monotonicity, IIA, and non-imposition. You can legislatively demand "the perfect voting system", but the universe is not obliged to ensure the existence of such a thing. Same for some types of robust ML, and no one knows which side of an impossibility theorem some precise-enough-to-code notion of political neutrality might fall on.

Which also brings up the REAL question: are tweaks to recommendation algorithms allowed? Obviously we can't ask FB/Twitter to freeze their recommendation algos -- it's their core product. So. If they notice an "obvious bias" and tweak the algorithm to correct for it, who decides whether that was a biased human intervention or a totally appropriate bug fix? Oh, right, a politically appointed FTC.

I think that "politically neutral" is impossible to formalize in code because it is a fundamental contradiction in terms. But even if it does, I suspect that any reasonable lists of formal specifications might be either mathematically impossible to train a classifier to respect or else at least AGI-complete to actually implement. But if you disagree, I'm happy to clone the Github repo and mess around with your proposal.

[jcranmer]

Note that the test actually proposed (not the press release blurb) says:

> The moderation practices of a provider of interactive computer services are politically biased if the provider moderates information provided by information content providers in a manner that [...] disproportionately restricts or promotes access to, or the availability of, information from a political party, political candidate, or political viewpoint

That means that any service that chooses to do something like suppress known conspiracy theories is going to fall afoul of the proposed changes.

[JoshTriplett]

For that matter, a policy of restricting hate speech will currently restrict one party more than another. A policy of prohibiting disinformation that could lead to voter suppression will currently restrict one party more than another. A policy of prohibiting misinformation about the ongoing pandemic will currently restrict one party more than another.

[untog]

> Showing "their algorithms and content-removal practices are politically neutral" is not an insurmountable bar.

It is when political appointees are the ones who judge if you've cleared the bar.

[SpicyLemonZest]

It seems pretty insurmountable to me. Can you go into more detail about how they'd do it? I've seen a lot of fights where one side says "putting this post up is biased against me" and another says "taking this post down is biased against me", and I'm not sure how Facebook could resolve those disputes with confidence the FCC won't say they did it wrong.

[jcranmer]

I think that's the point of this bill. They force Facebook et al. to get certified bias-free in a manner that basically makes it impossible to get that certification. So then they get the headlines saying "FCC finds Facebook is biased!!!111"

The bill would be more palatable to me if they simply dropped the immunity, without any certification process. But then demonstrating bias would require winning civil lawsuits, which requires demonstrating damage suffered by the bias and also convincing 12 members of the jury in a unanimous vote... which is unlikely to happen, I think.

(Addendum: actually, the real point of the bill may be to just say "Facebook/Twitter/Google is biased, and I'm doing something about it!" and ignore any actual chance of it making law or being reasonable. It's not like many people actually read details of bills to understand what it does and doesn't say.)

[throwawaygh]

> The bill would be more palatable to me if they simply dropped the immunity, without any certification process. But then demonstrating bias would require winning civil lawsuits, which requires demonstrating damage suffered by the bias and also convincing 12 members of the jury in a unanimous vote... which is unlikely to happen, I think.

No. Killing 230 entirely would allow Twitter and Facebook to be as politically biased as they want.

However, if one of their users libels you, then you could sue Facebook in addition to that user.

And if any Facebook user posts child porn, even for a short period of time, relevant parties at Facebook could face criminal charges for distribution.

You couldn't sue Facebook for being politically biased, but Facebook would be responsible for actual crimes that its users commit.

Hawley's bill says "you won't be responsible for the illegal stuff your users do (i.e., you get 230 protections), but only as long as you keep my political appointees happy."

[klyrs]

What does "politically neutral" mean, anyway? What happens if I establish a political party whose solitary goal is to torture babies, kittens and puppies to death? Is that suddenly a "political opinion" which must be protected?

[bosswipe]

The hypocrisy of the same FCC that said that network neutrality was too much regulation to now demand "political neutrality" is outrageous.

[heavyset_go]

> I don't think it would be world ending to get rid of Section 230

If you're running a start up, how would you feel knowing that if a user uploaded illegal content to your servers, you could be raided in the middle of the night and imprisoned for it?

Only those with billions of dollars to throw at moderation would be able to comply with the law. Everyone else would need to block user content by necessity, or risk having their lives ruined by malicious users.

The net result is that hosting free speech on the internet would be too risky for anyone other than giant corporations. The liability to host users' speech would be far too high for anyone else.

[vharuck]

Definitely. If 230 gets repealed, and somebody who had a forum wants to keep running it, they might consult a lawyer for advice. And that lawyer would say, "Don't allow user-created content. It's not worth the risk."

It only makes sense if the user content is the profit-generator and the forum owner ran the numbers and expects to still be profitable even after lawsuits.

So no more hobby forums, YouTube comments (some are good), or internet access in libraries:

>Kathleen R. v. City of Livermore, 87 Cal. App. 4th 684, 692 (2001).[136] The California Court of Appeal upheld the immunity of a city from claims of waste of public funds, nuisance, premises liability, and denial of substantive due process. The plaintiff's child downloaded pornography from a public library's computers, which did not restrict access to minors. The court found the library was not responsible for the content of the internet and explicitly found that section 230(c)(1) immunity covers governmental entities and taxpayer causes of action.

https://en.wikipedia.org/wiki/Section_230

[rrobukef]

Interesting. However, any manual action (choosing trusted posters, maintained database) is bad for adoption. Facebook and twitter take care of it, you should too.

Perhaps you should use karma and comment interactions to automatically attest the people you interact with. Add a "report" button to disavow certain users. Now there is a positive and negative feedback loop to reduce the workload of attestation.

Caveat: attestation must be stabilized. The existing hierarchies of admin/(super-)moderator work well as trusted posters. On the other hand, picking and choosing your moderator(s) is interesting and will birth new flame-wars and division.

Caveat (2): Adding more crypto explodes the amount of data which must be handled. Especially when every comment and upvote is signed.

[makomk]

What makes Section 230 a complicated and contentious issue isn't the actual details of the law - as you say, that's quite simple - it's the consequences of such a broad, powerful, simple, thing as protecting "interactive computer services" from almost all kinds of legal action for content created by others that they keep up, regardless of what they remove, with few caveats, across a vast swathe of causes for action, business models, moderation policies, etc.

For example, suppose you're an online service Twitbook used by a vast swathe of the world to communicate, and you decide that you want to allow calls to murder politicians you dislike but not (obviously) ones you like. Section 230 gives you pretty good protection from liability over your decisions as to which political figures get threatened with murder. Probably even if one of your users gets inspired and puts a bullet in the head of someone you'd like to see dead.

Or suppose you've got a nice legalized extortion racket seeking out negative claims about people or businesses, getting them to rank highly in Google, not allowing the original posters to remove them, and demanding money from the targets to take them down. Section 230 offers pretty much ironclad protection for your business model by making it nearly impossible to get a court order forcing you to take the content down, meaning you can ensure the only way to make it go away is to pay up, and you can even literally call the fee a charge to remove libellous or defamatory content and there's not a damn thing the court system will do about it. There's a long-running website Ripoff Report that has this as their business model, and they've won every case trying to get them to remove defamatory content without paying them money for the privilege thanks to Section 230. There's also plenty of imitators going after individuals, seeking out (say) claims they've cheated on their partners and charging money to remove them - again, solidly protected by Section 230.

[curryst]

> For example, suppose you're an online service Twitbook used by a vast swathe of the world to communicate, and you decide that you want to allow calls to murder politicians you dislike but not (obviously) ones you like. Section 230 gives you pretty good protection from liability over your decisions as to which political figures get threatened with murder

That's not true on 2 fronts. First, Section 230 requires good faith. That would almost certainly fail to pass the good faith muster, assuming they can demonstrate that it was done intentionally. So civilly, they would likely still be liable. In addition, Section 230 has no bearing on criminal law (it's specifically called out in subsection e). So in the event someone was killed, there would likely be a host of people from Twitter facing charges for being complicit in the death. They are effectively Charles Manson in this scenario, and I think they would have a hard time arguing that selectively filtering messages to expose users to messages encouraging them to kill someone does not count as speech.

I don't see why the second is a terrible issue. They're effectively a tabloid at that point, well known for spreading libelous content. I would be surprised if Queen Elizabeth is overly concerned that the tabloids say she's a lizard person. And also, RipoffReport is the wrong person to sue here, which is why that isn't working. If the content is libelous and you want it taken down, sue the person who wrote it, and have the judge issue a takedown order to Ripoff Report. Section 230 only protects them from civil liability, it doesn't make them immune to takedown requests.

In a funny idea, I wonder if you could upload a copyrighted image and then file a DMCA request against the page and have it delisted by Google. Their terms say you grant them a copyright, but if you upload a work that you don't own the copyright for you can't give them a copyright. Technically you're violating DMCA for the upload, and again by lying on the DMCA form you fill out (since you have to own the copyright) but as long as you pick something nobody is likely to sue you for, it should be fine (copy the credits from a book or something). Or if you want to get clever, you could have a friend make a painting of a stick figure in Paint and slap a copyright logo on it, then upload it and have your friend file the DMCA complaint. For bonus points, do it to every single page. They aren't liable because of section 230, but you could probably still force them to play a game of whack-a-mole with Google.

[gowld]

Why wasn't backpage.com projected by section 230?

[DanBC]

They were, until the owners of backpage started giving advice to child traffickers about how to continue advertising.

A child trafficker would place an ad featuring an image of child sexual abuse, with wording that gave coded hints that this was a child. "Amber Alert!"

Backpage would strip out that coded language and run the ad, with the image of child sexual abuse.

Sometimes those children would, after they'd been rescued, recognise themselves in the ads and ask backpage to take the ads down. Backpage refused.

https://www.justice.gov/file/1050276/download

[makomk]

I think the short answer is that backpage.com possibly was protected by Section 230, but federal and state prosecutors kept on harassing them with a barrage of new lawsuits and charges until they found a court willing to say it wasn't protected, and the people running it just gave in and pled guilty in the end. With the site shut down and no end in sight they just didn't have the resources to fight it.

[curryst]

e(5) of Section 230 is a specific carve out that the protections don't apply to sex trafficking content. I believe they used that to argue that Backpage wasn't protected by section 230.

[timwells3]

No, SESTA became law years after backpage was shutdown

[gowld]

Are anonymous trolls "information content providers"?

An easy fix is to say that an "information content provider" must be a legal person who is liable for their content. Then it's easy to find where the buck stops for a Tweet or a Rip-off Report or a Revenge Porn.

[incompatible]

Every HN profile must include a real name, verified by legal documents? I think they do that already in China.