I think this is a mistaken understanding of the ruling. In this case, TikTok decided, with no other context, to make a personalized recommendation to a user who visited their recommendation page. On HN, your front page is not different from my front page. (Indeed, there is no personalized recommendation page on HN, as far as I'm aware.)
I don't see how this is about personalization. HN has an algorithm that shows what it wants in the way it wants.
> TikTok, Inc., via its algorithm, recommended and promoted videos posted by third parties to ten-year-old Nylah Anderson on her uniquely curated “For You Page.”
Winding up in the same place as the defendant would require making a unique, dynamic, individualized BB for each user tailored to them based on pervasive online surveillance and the platform's own editorial "secret sauce."
Although HN's algorithm depends (mostly) on user input for how it presents the posts, it still favours some over others and still runs afoul here. You would need a literal 'most recent' chronological view and HN doesn't have that for comments. It probably should anyway!
@dang We need the option to view comments chronologically, please
That said, the feature you requested is already implemented but you have to know it is there. Dang mentioned it in a recent comment that I bookmarked: https://news.ycombinator.com/item?id=41230703
To see comments on this story sorted newest-first, change the link to
https://news.ycombinator.com/latest?id=41391868
instead of
I don't think the feature was that unknown. Per Wikipedia, the CDA passed in 1996 and Slashdot was created in 1997, and I doubt the latter's moderation/voting system was that unique.
You might like this then: https://hckrnews.com/
Where is the algorithmic recommandation that differs per user on HN?
In other words, the Supreme Court decision mentions editorial decisions but no court case has yet backed up if that means editorial decisions in the HN front page sense (as in mods make some choices but it’s not personalized.) Common sense may say mods making decisions is editorial decisions but it’s a gray area until a court case makes it clear. Precedence is the most important thing when interpreting law, and the only precedence we have is that it pertains to personalized feeds.
If you child porn that gets shared via direct links then that is bad even if nobody can see it, but it is much much worse if you start recommending that to people as well.
It's rarely if ever going to be the case that there is some distinct unit of code called "the algorithm" that can be separated and considered legally distinct from the rest of the codebase.
I don't think timestamps are, in any way, construed editorial judgement. They are a content agnostic related attribute.
The FYP algorithm that's contested in the case surfaced the video to the minor without her searching for that topic, following any specific content creator, or positively interacting (liking/favoriting/upvoting) with previous instances of said content. It was fed to her based on a combination of what TikTok knew about her demographic information, what was trending on the platform, and TikTok's editorial secret sauce. TikTok's algorithm made an active decision to surface this content to her, despite knowing that other children had died from similar challenge videos, they promoted it and should be liable for that promotion.
"(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"
Algorithmic ranking, curation, and promotion are not.
And they probably didn't think through how this particular argument could affect other areas of their business.
Importantly, this does not mean that before the Third Circuit decision platforms could just curate any which way they want and government couldn't regulate at all -- the mandatory removal regime around CSAM content is a great example of government regulating speech and forcing platforms to comply.
The Third Circuit decision, in a nutshell, is telling the platforms that they can't have their cake and eat it too. If they want to claim that their algorithmic feeds are speech that is protected from most government regulation, they can't simultaneously claim that these same algorithmic feeds are mere passive vessels for the speech of third parties. If that were the case, then their algorithms would enjoy no 1A protection from government regulation. (The content itself would still have 1A protection based on the rights of the creators, but the curation/ranking/privileging aspect would not).
This ruling is a natural consequence of the NetChoice ruling. Social media companies can't have it both ways.
> If that were the case, then their algorithms would enjoy no 1A protection from government regulation.
Well, the companies can still probably claim some 1st Amendment protections for their recommendation algorithms (for example, a law banning algorithmic political bias would be unconstitutional). All this ruling does is strip away the safe harbour protections, which weren't derived from the 1A in the first place.
Would it? The TV channels of old were heavily regulated well past 1st amendment limits.
Cable was never regulated like that. The medium actually mattered in this case
But the basis of this ruling is basically "well the Moody case says that curation/moderation/suggestion/whatever is First Amendment protected speech, therefore that's your speech and not somebody else's and so 230 doesn't apply and you can be liable for it." That rationale extends to basically any form of moderation or selection, personalized or not, and would blow a big hole in 230's protections.
Given generalized anti-Big-Tech sentiment on both ends of the political spectrum, I could see something that claimed to carve out just algorithmic personalization/suggestion from protection meeting with success, either out of the courts or Congress, but it really doesn't match the current law.
I see a lot of people saying this is a bad decision because it will have consequences they don't like, but the logic of the decision seems pretty damn airtight as you describe it. If the recommendation systems and moderation policies are the company's speech, then the company can be liable when the company "says", by way of their algorithmic "speech", to children that they should engage in some reckless activity likely to cause their death.
I even think it would count if the only moderation you did on your Lionel model train site was to make sure that most of the conversation was about Lionel model trains, and that they be treated in a positive (or at least neutral) manner. That degree of moderation, for that purpose, would make you liable if you left illegal or tortious content up i.e. if you moderate, you're a moderator, and your first duty is legal.
If you're just a dumb pipe, however, you're a dumb pipe and get section 230.
I wonder how this works with recommendation algorithms, though, seeing as they're also trade secrets. Even when they're not dark and predatory (advertising related.) If one has a recommendation algo that makes better e.g. song recommendations, you don't want to have to share it. Would it be something you'd have to privately reveal to a government agency (like having to reveal the composition of your fracking fluid to the EPA, as an example), and they would judge whether or not it was "editorial" or not?
[edit: that being said, it would probably be very hard to break the law with a song recommendation algorithm. But I'm sure you could run afoul of some financial law still on the books about payola, etc.]
I'm not sure that's quite it. As I read the article and think about its application to Tiktok, the problem was more that "the algorithm" was engaged in active and allegedly expressive promotion of the unsafe material. If a site like HN just doesn't remove bad content, then the residual promotion is not exactly Hacker News's expression, but rather its users'.
The situation might change if a liability-causing article were itself given 'second chance' promotion or another editorial thumb on the scale, but I certainly hope that such editorial management is done with enough care to practically avoid that case.
Personalisation simply matches users with the content the algorithm thinks they want to see. Moderation (which is typically also an algorithm) tries to remove harmful content from the platform altogether.
The ruling isn't saying that Section 230 doesn't apply because TikTok moderated. It's saying Section 230 doesn't apply because TikTok personalised, allegedly knew about the harmful content and allegedly didn't take enough action to moderate this harmful content.
These algorithms aren't matching you with what you want to see, they're trying to maximize your engagement- or, its what the operator wants you to see, so you'll use the site more and generate more data or revenue. Its a fine, but extremely important distinction.
If it does any customization based on what it knows about you, or what it tries to sell you because you are you, then it would be liable.
Yep., recommendation engines would have to be very carefully tuned, or you risk becoming liable. Recommending only curated content would be a way to protect yourself, but that costs money that companies don't have to pay today. It would be doable.
An account otherwise identical made two days later is going to interact with a different stream. Technically deterministic but in practice no two end up ever being exactly alike, (despite similar people having similar channels.)
The "answer" will turn back into tv channels. Have communities curate playlists of videos, and then anyone can go watch the playlist at any time. Reinvent broadcast tv / the subreddit.
This could very well be true for TikTok. Of course "selection" would include liked videos, how long you spend watching each video, and how many videos you have posted
And on the flip side a button that brings you to a random video would supply different content to users regardless of "selections".
You can't be licensing user content to AI as it's not yours. You also can't be undeleting posts people make (otherwise it's really reddit's posts and not theirs).
When you start treating user data as your own; it should become your own and that erodes 230.
undeleting is bad enough, but they've edited the content of user's comments too.
It is theirs. Users agreed to grant Reddit a license to use the content when they accepted the terms of service.
That kind of sounds... great? The only instance where I genuinely like to have a recommendation engine around is music steaming. Like yeah sometimes it does recommend great stuff. But anywhere else? No thank you
Ranking: A group of people share a ouija board, and together make selections.
Suggestion: A singular entity clips together media to create a new narrative, akin to a ransom note.
If the sum of the collection of content more than its parts, if is different in strength not kind, or self reenforcing, it's really hard to distinguish where the algorithm ends and the voters begin.
> "What does all this mean for Anderson’s claims? Well, § 230(c)(1)’s preemption of traditional publisher liability precludes Anderson from holding TikTok liable for the Blackout Challenge videos’ mere presence on TikTok’s platform. A conclusion Anderson’s counsel all but concedes. But § 230(c)(1) does not preempt distributor liability, so Anderson’s claims seeking to hold TikTok liable for continuing to host the Blackout Challenge videos knowing they were causing the death of children can proceed."
As-in, Dang would be liable if say somebody started a blackout challenge post on HN and he didn't start censoring all of them once news reports of programmers dieing broke out.
https://fingfx.thomsonreuters.com/gfx/legaldocs/mopaqabzypa/...
In general, judges will be ultimately responsible for evaluating whether "any", "sufficient", "appropriate", etc. actions were taken in each future case judgement they make. As with all things legalese, it's impossible to define with certainty a specific degree of action that is the uniform boundary of acceptable; but, as evident here, "none" is no longer permissible in that set.
(I am not your lawyer, this is not legal advice.)
This has interesting higher-order effects on free speech. Let's apply the same ruling to vaccine misinformation, or the ability to organize protests on social media (which opponents will probably call riots if there are any injuries)
But I mean what do you expect from a group of judges that themselves have written they're moving away from precedent?
Does TikTok have preempt this category of videos in the future or simply respond promptly when notified such a video is posted to their system?
Do you think its reasonable for social media to send videos to people without considering how harmful they are?
Do you even think its reasonable for search engine to respond to a specific request for this information?
It’s still curated, and not entirely automatically. Does it make a difference whether it’s curated individually or not?
I think that is quite generous. I think it is a deliberate reinterpretation of what the order says. The order states that 230(c)(1) provides immunity for removing harmful content after being made aware of it, i.e., moderation.