The short and sweet of it is that we need a way to enshrine “reciprocity” into the legal code that allows us to treat firms from authoritarian nations pursuing whole-of-society strategies against us as fundamental threats or in the same manner as our firms in those countries are treated, depending on the specifics of what the firm is and does.
If that requires a Constitutional amendment to allow discrimination based on a firm’s ownership, so be it, it’d pass. But I’d prefer that legal precedent from the Cold War be drawn on.
I also don't understand why there isn't an obvious economic reciprocity argument here. 1) China doesn't let US social media firms operate in China (or at least "allows" them to operate but in ways incompatible with their business.) 2) China doesn't let any business operate in China without partnering with a Chinese company and sharing IP.
Seems obvious that we shouldn't let TikTok operate US until FB/Insta/etc. are allowed in China.
(caveat: I don't actually know much about wrt Chinese import policy, so please correct me if I got it wrong.)
Yes, we're an open society and they're a closed society. Imitating authoritarianism is uh not going to be a real successful long-term policy for America. We are the wealthiest country in the world precisely because we don't block foreign competitors from our markets.
Not to mention- imagine if every democracy on Earth has TikTok, except for the US where it's banned. You can use TikTok in Britain, France, Germany, Italy, Canada, Japan, Mexico, Argentina.... but it's banned in the US. Imagine the propaganda coup for the CCP globally
I guess I was just thinking in comparison to what people in other countries are doing. Of course it would be a loss to current users, although I think that would go away pretty quickly, and after a couple of years people will have mostly forgotten (or moved on to something new).
This isn't about reciprocity, it's about Americans' First Amendment rights. During the height of the Cold War, the Supreme Court found that restricting the delivery of foreign Communist newspapers violated the right of Americans to receive and read those materials: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
No, that's the gist of it. There's a never-bloody-ending game of whack-a-mole whereby China tweaks policies to look as if it allows market access without ever actually doing so except in industries where a homegrown participant has already grown to present an unassailable network-effect moat, so the specifics change from industry to industry and literally year to year, but that's the overall picture.
What’s the economic benefit of allowing a Chinese company to operate a social media service in the US? China doesn’t have a comparative advantage here, they were just first to market for a product where network effects are everything.
Or to put another way, has china been economically helped or hindered by prohibiting US online tech companies to operate, and favoring their own (Baidu, etc), and why would the calculus be different for the US?
I guess if our user experience on a similar video app is worse, then tariff/banning them is sticking us with a worse home-grown version.
Network/first-mover effects are real but getting rid of them doesn't mean users would get an equivalently good version here (maybe they would if it were blocked)
Onto China specifics, or at least a few initial thoughts:
1. China's own regulations on social media, in addition to hamhanded censorship and a bunch of other stupid ideas, contain two useful policy kernels that should be considered elsewhere: deliberate attempts to constrain "virality" or at least confine its extent to harmless fluff, and some as yet very rough and overly wide-ranging attempts to constrain the ROIC of a bunch of software developers.
We absolutely do not want to mimic the regulatory and legal environment of the PRC in full, but as I state elsewhere, what's worth considering are regulatory attempts to stuff pandora back in the box. We should be looking for legal ways to take the major social media networks back to the circa 2005-08 model of "stay connected with people you know/follow a random smattering of the news of the day" We should also levy some pretty punitive taxes on social media companies. The product is free, so we cannot introduce a sin tax on the end user, but we can try to make capital steer clear of the sector. We'd thereby constrain its growth, regulatory capture, lobbying efforts, and attempts to get around our "rewind the clock" laws.
2. The CCP's thoughts and actions on the matter are very, very clear. Whenever and whereever they are able, they try to manipulate the information space of foreign countries, both in ways that redound to their own benefit and in ways that simply hurt powers they dislike. They have done this through social media, traditional print and broadcast media, education and joint research efforts, and outright threats, intimidation, and even extraordinary renditions of their own and other nations' citizens.
Taiwan, in particular, has been a target, as have Australia, South Korea, New Zealand, and the US.
TikTok cannot be permitted to remain in ByteDance's hands because ByteDance can and will immediately roll over for the CCP when asked. It will provide private information on users, tweak feed algorithms, censor content, anything the Party asks it to do. It already has done both the second and third in the past on the US and EU versions of its platform, and likely the first as well.
There are *no* legal protections for private firms in China to resist the government, and in fact the National Data Security Law put in place several years ago explicitly legalizes any Party attempts to extract data or force product changes on any tech. firm.
Given that there's probably a 1-in-5 chance that the US ends up in a shooting war with the PRC this century, no mainland Chinese firm should ever have a commanding position in the US media sphere. Period.
Whatever legal means can be found or created to make that happen must be pursued, and it has to happen soon.
One of the selling points of liberal democracy has long been that it doesn't need to engage in "constraining" the flow of information in order to survive or thrive. It seems wildly premature to be declaring that era over.
Far fewer constraints, perhaps. But I've never heard anyone except libertarians (see my other recent comment if you want my unvarnished opinion) say that liberal democracies do not need, would not benefit from, or cannot tolerate *any* constraints on information flow.
This "era" to which you refer never existed in the first place, so I'm not declaring anything over.
Current US and Chinese social media policies mainly differ in kind. Your plans by your own implicit admission would mean they merely differ in degree. I don't know what to call that other than the end of an era.
I grew up during the end of the cold war hearing all about how a free society like ours didn't need to engage in censorship. I don't think that was illusory or only a libertarian talking point. Your plans would certainly make it illusory though.
As I state elsewhere, there's no world in which *someone* isn't playing a role in deciding what you see and hear, the question is who, and secondarily, how accountable are they to you?
Right now the answer is "automated tools owned by for-profit enterprises motivated by maximizing engagement at all costs" and "not at all".
Based on that, I think it's necessary to start standing up and clipping the wings of *all* social media companies. It's just more urgent that we do it to the one owned by a geopolitical adversary first.
You're free to think of that what you will, but I'm not seeing any compelling counterarguments beyond just "that's unconstitutional". That may or may not be true, but it's certainly not anything more than an argument from authority.
No one has provided an argument that there's no real or potential harm in this, which is the only thing that's going to convince me that we don't need to do it.
Then you won't be convinced. You are not wrong about the potential for harm here. I'm sure we both think governments should be constrained in the lengths they can go in the pursuit of their goals, but we draw the line in different places. Your ideas cross that line for me.
I would like a fuller reconciliation between the views in this post with the views in the misinformation post that went public just yesterday (https://www.slowboring.com/p/misinformation?s=r).
There's a disconnect that I'll simplistically and churlishly state as follows: "Misinformation is a made-up issue, stop with that nonsense. Except for TikTok, that one really is misinformation and is bad."
I was thinking that too while reading this. My interpretation is that what he's talking about here isn't "misinformation" - it's about using real information to rile people up and polarize them in problematic ways.
Yes, that's the thing that kills me with people who think that they've somehow found an exception to the First Amendment by saying the word "algorithm." The logic, as far as I can tell, is that freedom of speech only exists in so far as it's ineffectual, but once it might start really influencing people it's time to crack down. It's like an unironic speech-related version of the (apocryphal?) Emma Goldman quote, "If voting changed anything, it would be made illegal."
This leads to my favorite meta point: I think some past SB articles need revisiting/rethinking and continued idea battles. I think it's great to start a (less stupid) discussion on this topic but this is one tough nut to crack.
Matt also needs to reconcile this with his (correct) understanding of why the Fairness Doctrine and other broadcasting related principles don't apply to the internet. If the FCC can't regulate content on-line, why would Matt think the US government can ban foreign content on-line?
Maybe one place to start would be drop the animus to China, which seems to be a growing theme on SB. We have our issues; they have theirs. Imposing our system all over the world hasn't worked out well in the last fifty years, either for us or other societies. I've never watch a TikTok, but I'm all for it if young people like it, no matter how much they're being brainwashed.
My guess is he thinks a bipartisan agreement about misinformation (or really declaring any issue to be beyond the pale) is different than blaming your partisan woes on the voters being brainwashed
Yeah, but there really is a disconnect between yesterday's 'stop whining about misinformation, you've got to meet the voters where they are' on the one hand and today's 'where's my regulatory solution to TikTok' on the other? What is preventing America's world-leading social media techies from designing a platform that simply does the same stuff better than TikTok, and 'meeting the users where they are'?
There’s certainly some tension and I think coming one after the other he should have addressed it.
To your comment there are lots of potential replies. Our world leading companies have profit motive so it’s certainly possible for a foreign govt to ‘outcompete’ them with an inferior and biased product. Just like China can ‘outcompete’ our banks for development projects in Africa by choosing to finance low return projects. Not clear that’s actually happening here. There’s a related point about certain demos getting captured by specific social media products: Facebook has a core age group, instagram slightly different, tik tok different again. It may not make sense for tech to compete with any established player head on. Perhaps better to work on a VR social network that aims at current 15 and below idk. And separately from all that there’s a question about how effective tik tok propaganda really is. I do think he needs to write more about this and be clearer.
Matt, when I saw the headline I was very excited because I thought you would *finally* address the point that I've brought up to you on Twitter multiple times over the past two years whenever you talk about restricting TiKTok: there is a First Amendment right to consume foreign-created Communist propaganda. That was decided by the Supreme Court in Lamont v. Postmaster General, 381 US 301 (1965). (Here's a summary of the case: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general ) Instead, you've completely ignored the issue, using an analogy about FCC-regulation of ownership of a broadcaster.
To be blunt, the broadcaster analogy is wholly inapposite for reasons that I *know* you understand -- ownership and restrictions on content on broadcast facilities was subject to FCC regulation (and thus to things like the Fairness Doctrine, bans on indecency, etc.) because of them being considered to operate on "public airwaves," a concept that doesn't apply to the internet.
Further, the argument for regulating or banning TikTok is dramatically WORSE than the argument for the restrictions on receiving foreign-created Communist propaganda that were struck down in Lamont. In that case, the issue was receiving copies of newspapers from Communist regimes. The content of those papers was obviously 100% created and selected by the Communist regimes and very little, if any of it, in any particular issue of the paper was created by American citizens. With TikToK, almost certainly the vast majority of the content on it is *not* created by the Chinese government and quite a bit of it is created by private American citizens. To put it another way, 100% of the content of Tass, People's Daily, etc. was generated by a Communist regime and selected for propaganda value, and there was no opportunity for anyone else to put any content in the newspaper. With TikTok, you have effectively an infinitely large "newspaper" that any user can contribute content to (including American citizens) and then the Chinese regime filters it pursuant to the TOS and chooses what order to display it in. While that's not the New England town square style of free speech, I can guarantee you there's dramatically more freedom in what people are posting to TikTok these days than what anyone was getting to publish in a Soviet or Communist Chinese newspaper in the 1960s!
The Warren Court created all types of rights that "flowed from" constitutional rights rather than actually being based in the text of the Constitution. I agree that Lamont is relevant but I think it's a live question whether the current Court would hold the same. TikTok also does not have an equivalent claim to the mantle "the press" in the same way that Communist newspapers did.
(1) The present Court is, if anything, more pro-free speech than the Warren Court was. Morse v. Frederick, 551 U.S. 393 (2007) (the "Bong Hits 4 Jesus" case) is basically the only decision in the last 25 years or so whose ruling clearly went against the free speech side and, in light of the ruling in Mahanoy Area School District v. B.L., 594 U.S. ___ (2021) (the "cursing cheerleader" case), it seems plausible Morse would be decided differently if it was brought today (replacing Scalia, Morse majority, with Gorsuch, Mahanoy majority, would flip Morse to 5-4 for the free speech side; adding Kavanaugh, Mahanoy majority, in place of Kennedy, Morse majority, would make it 6-3, so even if Barrett would have ruled differently than Ginsburg in Morse, the free speech side would still win).
(2) I would say the First Amendment claims for Americans to access and use TikTok are actually substantially stronger than the Communist newspapers because, as I explained above, American citizens can upload content to the platform and also have some degree of control over what material they view on it.
1) I am very skeptical that one can place the justices along a spectrum based on how "pro" free speech they are and arrive at anything with real predictive or substantive value. And one would probably also need to take into account how deferential each justice is towards national security justifications.
2) You can make that claim but it remains true that newspapers receive special solicitude in the first amendment context because of the 1) existence of a direct historical allegory and 2) the express language in the First Amendment singling out "the press". Those considerations do not apply to TikTok. And any legislation targeted towards "the algorithm" could be crafted to preserve the ability of people to upload/share/access content using the platform. In that way, the content algorithm is what is being regulated, not the underlying speech. I do not think the First Amendment entitles citizens access to a particular proprietary algorithm, personally. But who knows.
I think the First Amendment very clearly entitles citizens access to content curated by a "particular proprietary algorithm" because otherwise you've literally taken the position that viewpoint discrimination is A-OK so long as the magic word "algorithm" is uttered. The contents of the New York Times, CNN, Fox News, Wall Street Journal, etc. are 100% determined by algorithm too -- it's just that their "algorithms" exist in the heads of their executives/owners and editors-in-chief rather than as lines of code.
I also think your effort to keep pleading that there's a special distinction between "the press" and other forms of media outlets that somehow subjects those other outlets to reduced First Amendment protection is really ignorant of trends in First Amendment jurisprudence over the last 50 years, which is in the direction of finding that legacy media *don't* qualify for special treatment relative to other media outlets. (And that's setting aside the point that the key right here is the right to read/hear/see, rather than the right to publish. The Soviet Union and PRC had no First Amendment right to publish the papers at issue in Lamont; Lamont had a First Amendment right to receive the papers he had ordered.)
Every 1st amendment case that I can think of in the past 10 or even 20 years which was granted cert by the court resulted, to my memory, in an expansion of or recognition of new territory covered by the 1st amendment, not support of restrictions.
Exactly right. I commented elsewhere here that in roughly the last 25 years the only SCOTUS decision I can think of going against the First Amendment side was Morse v. Frederick, 551 U.S. 393 (2007) (the "Bong Hits 4 Jesus" case) and the vitality of even that decision has to be seen as being limited in light of Mahanoy Area School District v. B.L., 594 U.S. ___ (2021) (the "cursing cheerleader" case).
How is TikTok not 'the press'? I'm pretty sure 'the press' at the time of ratifying the First Amendment included 'the press' whether printed in the US or in hostile foreign countries, like England, as well as anything from serious broadsheets covering debates in Congress to tales of lurid sex crimes. Is SB 'the press'? Not if TikTok isn't.
In addition to the "public airwaves" concern, the other major support for ownership & content regulation in broadcast radio & television is the physics imposed limited resource constraint: there is only so much available broadcast spectrum in a given physical geography and therefore the resource is limited and thus must be utilized to the maximal benefit of the people. The internet isn't limited that way.
Yes, limited spectrum is also at issue for broadcasting, but that was the rationale for why broadcast communications had to be made the "public airwaves" in the first place, so I viewed the limited spectrum issue as being encompassed by what I said here. (I was trying to slam that comment out as fast as possible before going to work.)
Are you saying that Matt's argument is wrong on the merits, or that he's correctly identified a problem for which there is no solution that passes constitutional muster? Because frankly, I'm pretty sick of pretending that stare decisis and sacred constitutional principles mean that nine unelected hacks get to override policy whenever they feel they can get away with it. The Constitution is not a suicide pact.
I guess I see it as being wrong on the merits *and* that there is no solution that passes constitutional muster? Even if I thought it was constitutionally permissible to do what Matt proposes, I would be opposed to it. I believe people should be allowed to say things I disagree with or that even are directly harmful to my interests so long as they aren't advocating for immediate criminal action against me.
It is not a suicide pact for a government to abstain from censorship. Tik Tok presents a problem, and we should look for solutions, but giving the government the power to block off sections of the internet should be off the table in a liberal democracy. Look for solutions elsewhere.
"Banning Tiktok" doesn't need to entail putting up a Great Firewall. How about we start by preventing Apple and Google from listing it in their app stores?
This is simply not the solution being proposed. There is nothing on TikTok that wouldn’t be allowed on a TikTok that is not an apparatus of the Chinese Communist Party’s espionage operation. We just don’t want the CCP to select what is seen in secret.
The main proposed solution I'm hearing is to take government actions for the purpose of preventing a disfavored organization from choosing what they publish to the public. That falls pretty squarely under the umbrella of censorship.
Wait, uh, what are we supposed to do about TikTok's ownership? Did that part get left out of the piece? The 'USSR buying a TV station' analogy doesn't work because China didn't buy anything here- they built a platform, and Americans are voluntarily using it. A closer analogy would be to the USSR broadcasting a cable news channel domestically, and Americans somehow picking up the signal.
We're an open society with some of the most robust free speech laws (and property) laws on the planet, we can't ban foreign platforms, and we probably can't make TikTok sell to an American company. There's literally no legal basis for any of this. Thompson's piece was terrible (he said that we should block TikTok on national security grounds, which would have been laughed out of court), and that was right around the time that I canceled my 5 year Stratechery membership. (Thompson became a Joe Rogan/Ben Shapiro fReE sPeEcH nut, except when he wanted to ban TikTok, then went back to 'content moderation is censorship').
I agree with some of the problems raised here, but again these are some of the problems that open societies face. To be constructive I guess if it could be proven that TikTok is amplifying Russian or Chinese propaganda the US could pressure its advertisers, which would probably change its behavior. But uh otherwise this piece is bad, and Thompson's piece was worse. Sad to see how quickly pundits get away from the values of a free society
Yes, thank you, as I noted elsewhere, Matt's analogy here is really absurd and it's bizarre that he picked it given that he's obviously familiar with the "public airwaves" concept and why the FCC doesn't have regulatory control over internet content. Further, the discourse on this subject is really deficient without acknowledging that the right of Americans to receive foreign communist propaganda was already litigated and ruled on in Lamont v. Postmaster General, 381 U.S. 301 (1965), summary here: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
On the merits we need to do this, and if it requires a Constitutional Amendment, that's well within reach on this one issue.
But I suspect the free speech/libertarian faction of the comment section is *wildly* overstating their certainty that legal precedent is on their side, and no such thing will be necessary once there's an administration willing to do it without the Trumpian grandstanding bullshit.
Is it just me or is the problem here actually algorithmic curation of content? Like, if TikTok used the Twitter model of old (reverse-chronological feed of everything the people you follow posted) this wouldn’t be a problem. And this is the same dynamic as causes trouble at Facebook, and YouTube, and Twitter, and wherever else. Perhaps the real solution here is to make curation via algorithm in general just not an appealing business model.
People also like money to be freely printed in the short-term, plenty of people like cocaine, no one likes taxes...
But we ignore them in the name of holding society together with duct tape and baling wire (the common good, in polite parlance). I think there's a very, very compelling case for holding Facebook, Twitter, and TikTok at gunpoint and forcing them back to 2008.
Oh, I didn't realize you were proposing a *regulatory* solution.
I think part of the problem is it would be extremely difficult to create a bright line between the algorithms that are bad and the ones that are fine. Even if you banned, e.g. machine learning algorithms for social media specifically, you have to define what counts as "machine learning" and what counts as "social media" and you'll just end up with people trying to game the legal definitions.
I think a better model is cigarettes. We've done pretty good at constraining cigarette use in the United States, which is kind of obvious when you travel to foreign countries that haven't put in any effort and it feels like every other person you meet is smoking a cigarette. We did that through society-wide messaging that cigarettes are bad for you, that using them was a moral failing, etc. We are at the stage with social media addiction where our society can't quite bring itself to admit it has a problem, and so it can't combat it. Sure, some culture writers will lament "wow we are all addicted to social media how terrifying" but no one actually takes it seriously as a moral imperative to quit. Once we can admit that it's a problem and a vice, and make it high status to be someone who does not spend much time on social media (rather than high status being a large social media following), we can rein this in to a large degree.
I think the real path is to apply absolutely *punitive* taxes to the profits of social media companies, such that instead of playing definitional whack-a-mole every week in court, we're sucking the vast majority of the capital that plays in this sector out of it and sending it elsewhere hunting for better returns.
The equivalent of a tobacco tax, which is why I bring it up now.
That reinforces the cultural side of things that you raise.
"I think part of the problem is it would be extremely difficult to create a bright line between the algorithms that are bad and the ones that are fine."
No, it's quite easy. If you run a website right now and allow users to post content on that site, then Section 230 immunizes you from any liability for their content. But how you choose which content to show your users is an editorial judgment; we should amend section 230 so that you are subject to classical libel law for search-engine type behavior (what Volokh (2021), "Treating Social Media Platforms Like Common Carriers?" calls the "recommendation function" of social media — see p. 408-12 (or p. 33-37 of https://escholarship.org/content/qt0bz3d1fx/qt0bz3d1fx.pdf)). If searching Google turns up unmarked unreliable information, then Google should be liable for it; if Twitter or Facebook feeds promote lies, they should be liable for it; and if Tiktok selectively displays videos with a reckless disregard for the truth, then Tiktok should be liable for it.
This incentivizes algorithms that are explainable (because they can be testified about in court), predictable (because how else can you explain them?), and promote truth (because truth is an absolute defense against libel). I'm not sure if this doesn't also expose the old "reverse-chronological" algorithms to excessive liability, but then the amendment could say something like "algorithms that do not select posts based on content shall enjoy the fair reporting privilege, and shall not be vulnerable to liability under this act". That, I think should suffice.
The algorithm writ large might also be a problem, but it's *really* a problem when a malicious state actor is in control of it. For better or worse, we can trust Facebook to be self-serving and maximize ad revenue with its algorithm, but TikTok can't even meet that low bar, and, from my time using it, they might have the most effective personalizing algorithm in all of social media.
This article is insufficiently boring in its take.
I work in this space (Content Moderation) and it's clear nobody who really does has been talked to.
"Would it be crazy to think they are putting their thumbs on the scale in favor of pro-Russian content?"
Yes.
ByteDance has US engineers and english moderators are not from China. Any demand like "censor Ukraine war would almost certainly immediately leak". Chinese ownership wouldn't survive, it's barely survived now and they haven't even done anything. That would be a very expensive one-time information control effort.
Some key points you are missing for a more boring explanation:
1) Human moderators / content labelers are mostly not American.
2) Human Video moderation is extra expensive because you have to watch the whole thing and extra hard to label right.
3) Read the details of the actual guardian article: The real objection is that it's NOT censoring "bad" Ukraine content. Study is highly misleading. (see below)
4) Chinese and US apps are effectively different apps (this is the case for pretty much everything).
5) You aren't seeing Chinese content moderations (no ghosts allowed). You ARE probably seeing everyone in the world not-China which != US/UK.
6) Different countries have radically different content standards especially for under 18. Some examples:
China: No ghosts. You can have up to 3 skeletons, but they have to be cute.
Germany: Naked people ok. Guns bad.
US: Naked people bad. Guns ok.
Saudi Arabia: different...
Ask yourself the boring question:
"Which country is any human moderator doing ENGLISH language content moderation / labeling likely to hail from?"
Answer: India.
Boring explanation:
(1) TikTok grew rapidly which typically means they didn't have time to localize moderation standards much beyond a China / not-China separation. And not China has a lot of variation that's hard to satisfy quickly and it's not just about the US.
(2) Not much human moderation / labeling going on.
(3) To the extent that it is: Indian Content Moderation has different cultural biases than US journalists / US Foreign policy types.
(4) Many US ByteDance engineers there likely feel active political censorship at scale (lets be clear that this is the ask) should be resisted even if there currently is a bit of "war hysteria" in the air.
Study Details:
The algorithm here is clearly just doing item-item similarity and returning things you like. If you watch a video all the way to the very end that means you REALLY REALLY like it. The researchers clearly set their "test" up to sucker gullible journalists by watching every video to the "very end".
It's as if Facebook let you click on the "like" button 10 times.
Basically they kept telling TikTok they REALLY REALLY like Ukraine "Disinfo" videos and were "surprised" it kept recommending similar videos.
Excellent comment. Thanks for this info. It is very hard to trust random takes on TikTok bias when you can shift the algorithm to show you anything you want by just watching more of something, as you describe.
This was a pretty bad piece, I usually expect better from Matt. I assume it is because Matt doesn't use TikTok. Imagine writing a long article about Twitter having only read other people's takes and studies on how often Ukraine misinformation was seen.
Under 30s probably have favorable views on Russia because they grew up in the '1980s called and want their foreign policy back' phase of liberal ideology. (https://www.youtube.com/watch?v=T1409sXBleg) The way the red scare and Cold War are taught in history classrooms and the politicization of views on Russia probably also contribute.
re: TikTok and the algorithm, I tend to agree that where the algorithms can be avoided, they should. I certainly hope Matt is doing his part and remembering to set his Twitter, Next-door, and Facebook to be chronological. This is a required step to making them not be hate machines, but the US platforms are tricky and like to periodically revert the setting (for engagement, obv).
The 'accountant' thing is a reference to a TikTok meme/trend: https://www.youtube.com/watch?v=MqdchLYNEPY Not a nefarious censorship thing. Something folks don't really seem to understand is that while FYP is heavily moderated for new recommendations, once you follow someone, TikTok will mix in content from the people you follow.
I agree that China shouldn't control a major US media platform, but acting like this is somehow China's fault is bizarre and gives a huge pass to US interests who caused this problem. Specifically Vine, and Twitter killing Vine, caused this problem by identifying a clear market need, then leaving it empty for nearly a decade. Something I wish people in the US understood is that we have a responsibility to be good at our jobs, compete vigorously, do good work, and make sure that we are putting forth competitive offerings such that we don't leave giant market gaps for Chinese competitors.
Imagine clutching pearls over China when US tech companies are 100% responsible for this situation. It's not the vibe.
There was an opportunity to intervene before TikTok became a big deal (think spring/summer 2019) but of course the US missed that. Had US leadership been effective, they could have gone to YouTube and Instagram and said, "Can you please make Reels/Shorts competitive ASAP and not wait until TikTok hits critical mass?" https://www.nytimes.com/2019/03/10/style/what-is-tik-tok.html
Perhaps we could have also asked Facebook/Instagram/Twitter to make their algorithms not absolute dumpster fires of conflict and hate. TikTok is competitive because TikTok made a competitive product, and I'd like to see more emphasis on making something better than TikTok (surely if you believe in western values you agree it is possible) rather than trying to have the US government pick winners and losers. Doomscrolling is not a thing on TikTok.
I don’t know if anyone else has mentioned this, but it seems the whole problem with TikTok is that it is a fabulous video platform which became wildly successful while video platforms created by American firms like Facebook have been utter shite. TikTok apparently excels at ease-of-use while American attempts at video apps were clumsy and lacking in features. What happened to American technological ingenuity?
Matt’s argument here (which I’ve seen in others like Ben Thompson) is absurd and hypocritical. Even if TikTok's algorithm is 100% controlled by the PRC government, it competes freely and fairly in the open US market. It’s not a monopoly and has not abused its power in anti-competitive behaviors. Banning TikTok would be philosophically contradictory to not only our principles of free speech, but also that of market economy.
Matt’s argument is that we need to bubble wrap US citizens in US propaganda. Kevin Drum takes the next logical step and says that it needs to be center left bubble wrap and Fox News is the enemy. 1984 or a Brave New World or the animal farm book, one of those, had to warn that this is sort of a tricky subject.
I have a wholly different take, which is that we need to be educating our people from an early age to recognize propaganda and influence campaigns and to harden themselves from pernicious influence, to seek facts separate from opinion, to discern data from anecdote, to think critically. We can not coddle, we can only harden.
I can't shake the feeling that this is like a deeply problematic standard. Like everything about the specifics of the case seem right, but like we seize your firm based on very little actual evidence other than well they're Chinese and probably are doing something bad.
They made a fundamentally excellent product that fills a completely different niche than home grown socials and now it's going to be taken from them. It doesn't seem like fair play even if everything about the case is correct. What firms could you not destroy under the evidentiary standard of we think they’re probably doing something bad?
I'm sympathetic to this but China straight up bans American companies like Google and Facebook and to not do something about TikTok seems like unilateral disarmament.
I believe a lot of TikTok is now run from the US anyway and we could do something like have the US audit the algorithm, but that has it's own problems and may not even technically be possible.
I mean we know TT is doing all these things like we knew Iraq had a secret wmd program the inspectors couldn’t find. Like what could you not take under this kind of film flam everyone thinks it but no one has proved it standard?
The Soviet Union and PRC banned all sorts of American media during the Cold War, but that didn't stop Americans from having a First Amendment right to read media from those countries. See Lamont v. Postmaster General, 381 U.S. 301 (1965), summary here: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
If the objection is that TikTok is more effective at influencing Americans' attitudes and beliefs than Pravda was, then you've basically just invented the free speech version of the Emma Goldman quote about how "[i]f voting changed anything, it would be made illegal."
So because Chinese consumers are forced by their government to use lower quality services we should accept that for Americans on a just trust us basis?
I’m quite certain that if you extirpated any of these firms (Facebook, TikTok, YouTube) from the face of the earth tomorrow, any firm with sufficient server/AWS capacity could have a near-perfect replacement up and running within the year.
As for “just trust us”, we should we looking for a Cold War precedent sufficiently narrow to be reasonably confident it won’t be abused and then watching its use closely.
I'm not sure what is to be done about it, but TikTok's algo is *scary* good at personalizing and maximizing engagement. Use it for a week and then think about it for a minute and it's basically the social media tech equivalent of those fantastical GPT3 word models or DALL-E pictures out of OpenAI. The tech they've made is genuinely disruptive, and their growth reflects that. Compound on that that we have little idea what levers the CCP is pushing on the backend, and purely from a consequentialist perspective, I'm pretty convinced that it should have been banned under pretty much whatever spurious pretext the State department lawyers could have ginned up.
I've used it off and on for two years, and I don't find it find particularly good at personalizing, or particularly addictive. It's moderately entertaining, and I look at it once a month or two. I can see how it sort of personalizes over time, but honestly I don't see anything incredible or amazing there- I'd argue Youtube personalizes much better
lol, the main thing it does is just not factor the creator's past popularity or follower count into relevance scores so it's not as easily game-able by bots. Also, less ideal for "high importance" creators. You just aren't used to that.
Saying "we shouldn't let TikTok provide service to Americans" is equivalent to saying "Americans shouldn't be permitted to use TikTok". It's always worth considering that censorship/deplatforming/restricting access/etc. is imposing restrictions not just on the speaker/service, but the listener/user as well.
Obviously TikTok is doing censorship itself, and the idea here is to use the threat of banning it as leverage to get them to stop/sell to an American owner that maybe won't censor users, but cutting off access to a censored service is also a form of censorship.
The short and sweet of it is that we need a way to enshrine “reciprocity” into the legal code that allows us to treat firms from authoritarian nations pursuing whole-of-society strategies against us as fundamental threats or in the same manner as our firms in those countries are treated, depending on the specifics of what the firm is and does.
If that requires a Constitutional amendment to allow discrimination based on a firm’s ownership, so be it, it’d pass. But I’d prefer that legal precedent from the Cold War be drawn on.
More later on the China specifics below.
I also don't understand why there isn't an obvious economic reciprocity argument here. 1) China doesn't let US social media firms operate in China (or at least "allows" them to operate but in ways incompatible with their business.) 2) China doesn't let any business operate in China without partnering with a Chinese company and sharing IP.
Seems obvious that we shouldn't let TikTok operate US until FB/Insta/etc. are allowed in China.
(caveat: I don't actually know much about wrt Chinese import policy, so please correct me if I got it wrong.)
Yes, we're an open society and they're a closed society. Imitating authoritarianism is uh not going to be a real successful long-term policy for America. We are the wealthiest country in the world precisely because we don't block foreign competitors from our markets.
Not to mention- imagine if every democracy on Earth has TikTok, except for the US where it's banned. You can use TikTok in Britain, France, Germany, Italy, Canada, Japan, Mexico, Argentina.... but it's banned in the US. Imagine the propaganda coup for the CCP globally
I don't follow how that's a global propaganda coup?
Communist regimes have historically loved to rub hypocrisy on civil liberties issues in the US's face.
No one would be banning the idea of a tiktok-like app, just one based in/run by China. I assume a new American version would pop up pretty quickly.
Its called Vine. It failed.
I'm pretty sure it would just be "out of sight, out of mind".
After two years of apparently everyone being on TikTok 24/7 during the pandemic, it would be a bit harder to be properly "out of sight, out of mind".
I guess I was just thinking in comparison to what people in other countries are doing. Of course it would be a loss to current users, although I think that would go away pretty quickly, and after a couple of years people will have mostly forgotten (or moved on to something new).
This isn't about reciprocity, it's about Americans' First Amendment rights. During the height of the Cold War, the Supreme Court found that restricting the delivery of foreign Communist newspapers violated the right of Americans to receive and read those materials: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
No, that's the gist of it. There's a never-bloody-ending game of whack-a-mole whereby China tweaks policies to look as if it allows market access without ever actually doing so except in industries where a homegrown participant has already grown to present an unassailable network-effect moat, so the specifics change from industry to industry and literally year to year, but that's the overall picture.
Well I happen to think reciprocal tariffs are bad - even if they do a tariff, I don't think we should do this on economic grounds.
I'm fine with it on NatSec grounds though.
What’s the economic benefit of allowing a Chinese company to operate a social media service in the US? China doesn’t have a comparative advantage here, they were just first to market for a product where network effects are everything.
Or to put another way, has china been economically helped or hindered by prohibiting US online tech companies to operate, and favoring their own (Baidu, etc), and why would the calculus be different for the US?
Also I suspect they've been economically hindered but politically helped.
I guess if our user experience on a similar video app is worse, then tariff/banning them is sticking us with a worse home-grown version.
Network/first-mover effects are real but getting rid of them doesn't mean users would get an equivalently good version here (maybe they would if it were blocked)
Onto China specifics, or at least a few initial thoughts:
1. China's own regulations on social media, in addition to hamhanded censorship and a bunch of other stupid ideas, contain two useful policy kernels that should be considered elsewhere: deliberate attempts to constrain "virality" or at least confine its extent to harmless fluff, and some as yet very rough and overly wide-ranging attempts to constrain the ROIC of a bunch of software developers.
We absolutely do not want to mimic the regulatory and legal environment of the PRC in full, but as I state elsewhere, what's worth considering are regulatory attempts to stuff pandora back in the box. We should be looking for legal ways to take the major social media networks back to the circa 2005-08 model of "stay connected with people you know/follow a random smattering of the news of the day" We should also levy some pretty punitive taxes on social media companies. The product is free, so we cannot introduce a sin tax on the end user, but we can try to make capital steer clear of the sector. We'd thereby constrain its growth, regulatory capture, lobbying efforts, and attempts to get around our "rewind the clock" laws.
2. The CCP's thoughts and actions on the matter are very, very clear. Whenever and whereever they are able, they try to manipulate the information space of foreign countries, both in ways that redound to their own benefit and in ways that simply hurt powers they dislike. They have done this through social media, traditional print and broadcast media, education and joint research efforts, and outright threats, intimidation, and even extraordinary renditions of their own and other nations' citizens.
Taiwan, in particular, has been a target, as have Australia, South Korea, New Zealand, and the US.
TikTok cannot be permitted to remain in ByteDance's hands because ByteDance can and will immediately roll over for the CCP when asked. It will provide private information on users, tweak feed algorithms, censor content, anything the Party asks it to do. It already has done both the second and third in the past on the US and EU versions of its platform, and likely the first as well.
There are *no* legal protections for private firms in China to resist the government, and in fact the National Data Security Law put in place several years ago explicitly legalizes any Party attempts to extract data or force product changes on any tech. firm.
Given that there's probably a 1-in-5 chance that the US ends up in a shooting war with the PRC this century, no mainland Chinese firm should ever have a commanding position in the US media sphere. Period.
Whatever legal means can be found or created to make that happen must be pursued, and it has to happen soon.
One of the selling points of liberal democracy has long been that it doesn't need to engage in "constraining" the flow of information in order to survive or thrive. It seems wildly premature to be declaring that era over.
Far fewer constraints, perhaps. But I've never heard anyone except libertarians (see my other recent comment if you want my unvarnished opinion) say that liberal democracies do not need, would not benefit from, or cannot tolerate *any* constraints on information flow.
This "era" to which you refer never existed in the first place, so I'm not declaring anything over.
Current US and Chinese social media policies mainly differ in kind. Your plans by your own implicit admission would mean they merely differ in degree. I don't know what to call that other than the end of an era.
I grew up during the end of the cold war hearing all about how a free society like ours didn't need to engage in censorship. I don't think that was illusory or only a libertarian talking point. Your plans would certainly make it illusory though.
As I state elsewhere, there's no world in which *someone* isn't playing a role in deciding what you see and hear, the question is who, and secondarily, how accountable are they to you?
Right now the answer is "automated tools owned by for-profit enterprises motivated by maximizing engagement at all costs" and "not at all".
Based on that, I think it's necessary to start standing up and clipping the wings of *all* social media companies. It's just more urgent that we do it to the one owned by a geopolitical adversary first.
You're free to think of that what you will, but I'm not seeing any compelling counterarguments beyond just "that's unconstitutional". That may or may not be true, but it's certainly not anything more than an argument from authority.
No one has provided an argument that there's no real or potential harm in this, which is the only thing that's going to convince me that we don't need to do it.
Then you won't be convinced. You are not wrong about the potential for harm here. I'm sure we both think governments should be constrained in the lengths they can go in the pursuit of their goals, but we draw the line in different places. Your ideas cross that line for me.
Is SubStack a social media company? If not, why not?
Did anyone else not get the email? (Yes I checked spam)
I'm up early most days feeding an infant and need my newsletters!
The algorithm must have blocked it.
The email just went out, and there was rejoicing throughout the land.
I didn't get it either.
Didn’t get it and this post was GREAT. Glad I checked the site
I did not get the email. And no, I haven’t downloaded the Substack app.
Same. Was kinda worried ngl.
I didn't get it either - Just happened to see it in my Twitter feed (which I visit about once a week so that is a little weird too.)
Wild night, last night.
I didn’t get the email either
Same.
I did not get the email either.
Not me
I get the SB daily article email just after 6am every day... I was getting worried for the SB team's health and safety! ;)
My situation is similar to yours except that I never seem to get an email on Thursday
Unfortunately, no Android app yet.
I have not. Should I? I don't know how that relates to the email version.
oh that's awesome! I'll look into it. Thanks.
I would like a fuller reconciliation between the views in this post with the views in the misinformation post that went public just yesterday (https://www.slowboring.com/p/misinformation?s=r).
There's a disconnect that I'll simplistically and churlishly state as follows: "Misinformation is a made-up issue, stop with that nonsense. Except for TikTok, that one really is misinformation and is bad."
I was thinking that too while reading this. My interpretation is that what he's talking about here isn't "misinformation" - it's about using real information to rile people up and polarize them in problematic ways.
Yes, that's the thing that kills me with people who think that they've somehow found an exception to the First Amendment by saying the word "algorithm." The logic, as far as I can tell, is that freedom of speech only exists in so far as it's ineffectual, but once it might start really influencing people it's time to crack down. It's like an unironic speech-related version of the (apocryphal?) Emma Goldman quote, "If voting changed anything, it would be made illegal."
This leads to my favorite meta point: I think some past SB articles need revisiting/rethinking and continued idea battles. I think it's great to start a (less stupid) discussion on this topic but this is one tough nut to crack.
This is why I had an issue with yesterday's post conflating "disinformation" with "misinformation."
Matt also needs to reconcile this with his (correct) understanding of why the Fairness Doctrine and other broadcasting related principles don't apply to the internet. If the FCC can't regulate content on-line, why would Matt think the US government can ban foreign content on-line?
Maybe one place to start would be drop the animus to China, which seems to be a growing theme on SB. We have our issues; they have theirs. Imposing our system all over the world hasn't worked out well in the last fifty years, either for us or other societies. I've never watch a TikTok, but I'm all for it if young people like it, no matter how much they're being brainwashed.
My guess is he thinks a bipartisan agreement about misinformation (or really declaring any issue to be beyond the pale) is different than blaming your partisan woes on the voters being brainwashed
Yeah, but there really is a disconnect between yesterday's 'stop whining about misinformation, you've got to meet the voters where they are' on the one hand and today's 'where's my regulatory solution to TikTok' on the other? What is preventing America's world-leading social media techies from designing a platform that simply does the same stuff better than TikTok, and 'meeting the users where they are'?
There’s certainly some tension and I think coming one after the other he should have addressed it.
To your comment there are lots of potential replies. Our world leading companies have profit motive so it’s certainly possible for a foreign govt to ‘outcompete’ them with an inferior and biased product. Just like China can ‘outcompete’ our banks for development projects in Africa by choosing to finance low return projects. Not clear that’s actually happening here. There’s a related point about certain demos getting captured by specific social media products: Facebook has a core age group, instagram slightly different, tik tok different again. It may not make sense for tech to compete with any established player head on. Perhaps better to work on a VR social network that aims at current 15 and below idk. And separately from all that there’s a question about how effective tik tok propaganda really is. I do think he needs to write more about this and be clearer.
Matt, when I saw the headline I was very excited because I thought you would *finally* address the point that I've brought up to you on Twitter multiple times over the past two years whenever you talk about restricting TiKTok: there is a First Amendment right to consume foreign-created Communist propaganda. That was decided by the Supreme Court in Lamont v. Postmaster General, 381 US 301 (1965). (Here's a summary of the case: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general ) Instead, you've completely ignored the issue, using an analogy about FCC-regulation of ownership of a broadcaster.
To be blunt, the broadcaster analogy is wholly inapposite for reasons that I *know* you understand -- ownership and restrictions on content on broadcast facilities was subject to FCC regulation (and thus to things like the Fairness Doctrine, bans on indecency, etc.) because of them being considered to operate on "public airwaves," a concept that doesn't apply to the internet.
Further, the argument for regulating or banning TikTok is dramatically WORSE than the argument for the restrictions on receiving foreign-created Communist propaganda that were struck down in Lamont. In that case, the issue was receiving copies of newspapers from Communist regimes. The content of those papers was obviously 100% created and selected by the Communist regimes and very little, if any of it, in any particular issue of the paper was created by American citizens. With TikToK, almost certainly the vast majority of the content on it is *not* created by the Chinese government and quite a bit of it is created by private American citizens. To put it another way, 100% of the content of Tass, People's Daily, etc. was generated by a Communist regime and selected for propaganda value, and there was no opportunity for anyone else to put any content in the newspaper. With TikTok, you have effectively an infinitely large "newspaper" that any user can contribute content to (including American citizens) and then the Chinese regime filters it pursuant to the TOS and chooses what order to display it in. While that's not the New England town square style of free speech, I can guarantee you there's dramatically more freedom in what people are posting to TikTok these days than what anyone was getting to publish in a Soviet or Communist Chinese newspaper in the 1960s!
The Warren Court created all types of rights that "flowed from" constitutional rights rather than actually being based in the text of the Constitution. I agree that Lamont is relevant but I think it's a live question whether the current Court would hold the same. TikTok also does not have an equivalent claim to the mantle "the press" in the same way that Communist newspapers did.
(1) The present Court is, if anything, more pro-free speech than the Warren Court was. Morse v. Frederick, 551 U.S. 393 (2007) (the "Bong Hits 4 Jesus" case) is basically the only decision in the last 25 years or so whose ruling clearly went against the free speech side and, in light of the ruling in Mahanoy Area School District v. B.L., 594 U.S. ___ (2021) (the "cursing cheerleader" case), it seems plausible Morse would be decided differently if it was brought today (replacing Scalia, Morse majority, with Gorsuch, Mahanoy majority, would flip Morse to 5-4 for the free speech side; adding Kavanaugh, Mahanoy majority, in place of Kennedy, Morse majority, would make it 6-3, so even if Barrett would have ruled differently than Ginsburg in Morse, the free speech side would still win).
(2) I would say the First Amendment claims for Americans to access and use TikTok are actually substantially stronger than the Communist newspapers because, as I explained above, American citizens can upload content to the platform and also have some degree of control over what material they view on it.
1) I am very skeptical that one can place the justices along a spectrum based on how "pro" free speech they are and arrive at anything with real predictive or substantive value. And one would probably also need to take into account how deferential each justice is towards national security justifications.
2) You can make that claim but it remains true that newspapers receive special solicitude in the first amendment context because of the 1) existence of a direct historical allegory and 2) the express language in the First Amendment singling out "the press". Those considerations do not apply to TikTok. And any legislation targeted towards "the algorithm" could be crafted to preserve the ability of people to upload/share/access content using the platform. In that way, the content algorithm is what is being regulated, not the underlying speech. I do not think the First Amendment entitles citizens access to a particular proprietary algorithm, personally. But who knows.
I think the First Amendment very clearly entitles citizens access to content curated by a "particular proprietary algorithm" because otherwise you've literally taken the position that viewpoint discrimination is A-OK so long as the magic word "algorithm" is uttered. The contents of the New York Times, CNN, Fox News, Wall Street Journal, etc. are 100% determined by algorithm too -- it's just that their "algorithms" exist in the heads of their executives/owners and editors-in-chief rather than as lines of code.
I also think your effort to keep pleading that there's a special distinction between "the press" and other forms of media outlets that somehow subjects those other outlets to reduced First Amendment protection is really ignorant of trends in First Amendment jurisprudence over the last 50 years, which is in the direction of finding that legacy media *don't* qualify for special treatment relative to other media outlets. (And that's setting aside the point that the key right here is the right to read/hear/see, rather than the right to publish. The Soviet Union and PRC had no First Amendment right to publish the papers at issue in Lamont; Lamont had a First Amendment right to receive the papers he had ordered.)
Every 1st amendment case that I can think of in the past 10 or even 20 years which was granted cert by the court resulted, to my memory, in an expansion of or recognition of new territory covered by the 1st amendment, not support of restrictions.
Exactly right. I commented elsewhere here that in roughly the last 25 years the only SCOTUS decision I can think of going against the First Amendment side was Morse v. Frederick, 551 U.S. 393 (2007) (the "Bong Hits 4 Jesus" case) and the vitality of even that decision has to be seen as being limited in light of Mahanoy Area School District v. B.L., 594 U.S. ___ (2021) (the "cursing cheerleader" case).
How is TikTok not 'the press'? I'm pretty sure 'the press' at the time of ratifying the First Amendment included 'the press' whether printed in the US or in hostile foreign countries, like England, as well as anything from serious broadsheets covering debates in Congress to tales of lurid sex crimes. Is SB 'the press'? Not if TikTok isn't.
In addition to the "public airwaves" concern, the other major support for ownership & content regulation in broadcast radio & television is the physics imposed limited resource constraint: there is only so much available broadcast spectrum in a given physical geography and therefore the resource is limited and thus must be utilized to the maximal benefit of the people. The internet isn't limited that way.
Yes, limited spectrum is also at issue for broadcasting, but that was the rationale for why broadcast communications had to be made the "public airwaves" in the first place, so I viewed the limited spectrum issue as being encompassed by what I said here. (I was trying to slam that comment out as fast as possible before going to work.)
Are you saying that Matt's argument is wrong on the merits, or that he's correctly identified a problem for which there is no solution that passes constitutional muster? Because frankly, I'm pretty sick of pretending that stare decisis and sacred constitutional principles mean that nine unelected hacks get to override policy whenever they feel they can get away with it. The Constitution is not a suicide pact.
I guess I see it as being wrong on the merits *and* that there is no solution that passes constitutional muster? Even if I thought it was constitutionally permissible to do what Matt proposes, I would be opposed to it. I believe people should be allowed to say things I disagree with or that even are directly harmful to my interests so long as they aren't advocating for immediate criminal action against me.
I have to say after reading Matt's post I was very rah-rah despite being very much a free speech person so thank you for providing perspective.
I'm still inclined to believe that it can bring real harms but that doesn't mean the (proposed) cure is better than the disease.
It is not a suicide pact for a government to abstain from censorship. Tik Tok presents a problem, and we should look for solutions, but giving the government the power to block off sections of the internet should be off the table in a liberal democracy. Look for solutions elsewhere.
"Banning Tiktok" doesn't need to entail putting up a Great Firewall. How about we start by preventing Apple and Google from listing it in their app stores?
And now we're back at very clear first amendment issues.
This is simply not the solution being proposed. There is nothing on TikTok that wouldn’t be allowed on a TikTok that is not an apparatus of the Chinese Communist Party’s espionage operation. We just don’t want the CCP to select what is seen in secret.
The main proposed solution I'm hearing is to take government actions for the purpose of preventing a disfavored organization from choosing what they publish to the public. That falls pretty squarely under the umbrella of censorship.
This was right up there with buying Greenland as "actually good Trump ideas"
Wait, uh, what are we supposed to do about TikTok's ownership? Did that part get left out of the piece? The 'USSR buying a TV station' analogy doesn't work because China didn't buy anything here- they built a platform, and Americans are voluntarily using it. A closer analogy would be to the USSR broadcasting a cable news channel domestically, and Americans somehow picking up the signal.
We're an open society with some of the most robust free speech laws (and property) laws on the planet, we can't ban foreign platforms, and we probably can't make TikTok sell to an American company. There's literally no legal basis for any of this. Thompson's piece was terrible (he said that we should block TikTok on national security grounds, which would have been laughed out of court), and that was right around the time that I canceled my 5 year Stratechery membership. (Thompson became a Joe Rogan/Ben Shapiro fReE sPeEcH nut, except when he wanted to ban TikTok, then went back to 'content moderation is censorship').
I agree with some of the problems raised here, but again these are some of the problems that open societies face. To be constructive I guess if it could be proven that TikTok is amplifying Russian or Chinese propaganda the US could pressure its advertisers, which would probably change its behavior. But uh otherwise this piece is bad, and Thompson's piece was worse. Sad to see how quickly pundits get away from the values of a free society
Yes, thank you, as I noted elsewhere, Matt's analogy here is really absurd and it's bizarre that he picked it given that he's obviously familiar with the "public airwaves" concept and why the FCC doesn't have regulatory control over internet content. Further, the discourse on this subject is really deficient without acknowledging that the right of Americans to receive foreign communist propaganda was already litigated and ruled on in Lamont v. Postmaster General, 381 U.S. 301 (1965), summary here: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
"No legal basis for any of this" -- Matt addresses this, up to the potential need for a constitutional amendment, and he's talking to lawmakers too.
The first question to answer is "is it a good idea on the merits." If so, it makes sense to pursue changes to the law to support it.
ah, I stupidly confused a comment here on the potential need for a constitutional amendment w/ Matt's article. Matt doesn't mention.
That was me, lol.
On the merits we need to do this, and if it requires a Constitutional Amendment, that's well within reach on this one issue.
But I suspect the free speech/libertarian faction of the comment section is *wildly* overstating their certainty that legal precedent is on their side, and no such thing will be necessary once there's an administration willing to do it without the Trumpian grandstanding bullshit.
I doubt an amendment is necessary. SCOTUS has a long-storied history of balancing rights that are in conflict with each other.
Is it just me or is the problem here actually algorithmic curation of content? Like, if TikTok used the Twitter model of old (reverse-chronological feed of everything the people you follow posted) this wouldn’t be a problem. And this is the same dynamic as causes trouble at Facebook, and YouTube, and Twitter, and wherever else. Perhaps the real solution here is to make curation via algorithm in general just not an appealing business model.
the ML-curated algorithms are an appealing business model because they provide a superior user experience to chronological order
i don't know how you could reasonably undo this. people like getting personalized content streams
People also like money to be freely printed in the short-term, plenty of people like cocaine, no one likes taxes...
But we ignore them in the name of holding society together with duct tape and baling wire (the common good, in polite parlance). I think there's a very, very compelling case for holding Facebook, Twitter, and TikTok at gunpoint and forcing them back to 2008.
How to do it legally, I have no idea.
Oh, I didn't realize you were proposing a *regulatory* solution.
I think part of the problem is it would be extremely difficult to create a bright line between the algorithms that are bad and the ones that are fine. Even if you banned, e.g. machine learning algorithms for social media specifically, you have to define what counts as "machine learning" and what counts as "social media" and you'll just end up with people trying to game the legal definitions.
I think a better model is cigarettes. We've done pretty good at constraining cigarette use in the United States, which is kind of obvious when you travel to foreign countries that haven't put in any effort and it feels like every other person you meet is smoking a cigarette. We did that through society-wide messaging that cigarettes are bad for you, that using them was a moral failing, etc. We are at the stage with social media addiction where our society can't quite bring itself to admit it has a problem, and so it can't combat it. Sure, some culture writers will lament "wow we are all addicted to social media how terrifying" but no one actually takes it seriously as a moral imperative to quit. Once we can admit that it's a problem and a vice, and make it high status to be someone who does not spend much time on social media (rather than high status being a large social media following), we can rein this in to a large degree.
I think the real path is to apply absolutely *punitive* taxes to the profits of social media companies, such that instead of playing definitional whack-a-mole every week in court, we're sucking the vast majority of the capital that plays in this sector out of it and sending it elsewhere hunting for better returns.
The equivalent of a tobacco tax, which is why I bring it up now.
That reinforces the cultural side of things that you raise.
"I think part of the problem is it would be extremely difficult to create a bright line between the algorithms that are bad and the ones that are fine."
No, it's quite easy. If you run a website right now and allow users to post content on that site, then Section 230 immunizes you from any liability for their content. But how you choose which content to show your users is an editorial judgment; we should amend section 230 so that you are subject to classical libel law for search-engine type behavior (what Volokh (2021), "Treating Social Media Platforms Like Common Carriers?" calls the "recommendation function" of social media — see p. 408-12 (or p. 33-37 of https://escholarship.org/content/qt0bz3d1fx/qt0bz3d1fx.pdf)). If searching Google turns up unmarked unreliable information, then Google should be liable for it; if Twitter or Facebook feeds promote lies, they should be liable for it; and if Tiktok selectively displays videos with a reckless disregard for the truth, then Tiktok should be liable for it.
This incentivizes algorithms that are explainable (because they can be testified about in court), predictable (because how else can you explain them?), and promote truth (because truth is an absolute defense against libel). I'm not sure if this doesn't also expose the old "reverse-chronological" algorithms to excessive liability, but then the amendment could say something like "algorithms that do not select posts based on content shall enjoy the fair reporting privilege, and shall not be vulnerable to liability under this act". That, I think should suffice.
The algorithm writ large might also be a problem, but it's *really* a problem when a malicious state actor is in control of it. For better or worse, we can trust Facebook to be self-serving and maximize ad revenue with its algorithm, but TikTok can't even meet that low bar, and, from my time using it, they might have the most effective personalizing algorithm in all of social media.
This article is insufficiently boring in its take.
I work in this space (Content Moderation) and it's clear nobody who really does has been talked to.
"Would it be crazy to think they are putting their thumbs on the scale in favor of pro-Russian content?"
Yes.
ByteDance has US engineers and english moderators are not from China. Any demand like "censor Ukraine war would almost certainly immediately leak". Chinese ownership wouldn't survive, it's barely survived now and they haven't even done anything. That would be a very expensive one-time information control effort.
Some key points you are missing for a more boring explanation:
1) Human moderators / content labelers are mostly not American.
2) Human Video moderation is extra expensive because you have to watch the whole thing and extra hard to label right.
3) Read the details of the actual guardian article: The real objection is that it's NOT censoring "bad" Ukraine content. Study is highly misleading. (see below)
4) Chinese and US apps are effectively different apps (this is the case for pretty much everything).
5) You aren't seeing Chinese content moderations (no ghosts allowed). You ARE probably seeing everyone in the world not-China which != US/UK.
6) Different countries have radically different content standards especially for under 18. Some examples:
China: No ghosts. You can have up to 3 skeletons, but they have to be cute.
Germany: Naked people ok. Guns bad.
US: Naked people bad. Guns ok.
Saudi Arabia: different...
Ask yourself the boring question:
"Which country is any human moderator doing ENGLISH language content moderation / labeling likely to hail from?"
Answer: India.
Boring explanation:
(1) TikTok grew rapidly which typically means they didn't have time to localize moderation standards much beyond a China / not-China separation. And not China has a lot of variation that's hard to satisfy quickly and it's not just about the US.
(2) Not much human moderation / labeling going on.
(3) To the extent that it is: Indian Content Moderation has different cultural biases than US journalists / US Foreign policy types.
(4) Many US ByteDance engineers there likely feel active political censorship at scale (lets be clear that this is the ask) should be resisted even if there currently is a bit of "war hysteria" in the air.
Study Details:
The algorithm here is clearly just doing item-item similarity and returning things you like. If you watch a video all the way to the very end that means you REALLY REALLY like it. The researchers clearly set their "test" up to sucker gullible journalists by watching every video to the "very end".
It's as if Facebook let you click on the "like" button 10 times.
Basically they kept telling TikTok they REALLY REALLY like Ukraine "Disinfo" videos and were "surprised" it kept recommending similar videos.
Excellent comment. Thanks for this info. It is very hard to trust random takes on TikTok bias when you can shift the algorithm to show you anything you want by just watching more of something, as you describe.
This was a pretty bad piece, I usually expect better from Matt. I assume it is because Matt doesn't use TikTok. Imagine writing a long article about Twitter having only read other people's takes and studies on how often Ukraine misinformation was seen.
Under 30s probably have favorable views on Russia because they grew up in the '1980s called and want their foreign policy back' phase of liberal ideology. (https://www.youtube.com/watch?v=T1409sXBleg) The way the red scare and Cold War are taught in history classrooms and the politicization of views on Russia probably also contribute.
re: TikTok and the algorithm, I tend to agree that where the algorithms can be avoided, they should. I certainly hope Matt is doing his part and remembering to set his Twitter, Next-door, and Facebook to be chronological. This is a required step to making them not be hate machines, but the US platforms are tricky and like to periodically revert the setting (for engagement, obv).
The 'accountant' thing is a reference to a TikTok meme/trend: https://www.youtube.com/watch?v=MqdchLYNEPY Not a nefarious censorship thing. Something folks don't really seem to understand is that while FYP is heavily moderated for new recommendations, once you follow someone, TikTok will mix in content from the people you follow.
I agree that China shouldn't control a major US media platform, but acting like this is somehow China's fault is bizarre and gives a huge pass to US interests who caused this problem. Specifically Vine, and Twitter killing Vine, caused this problem by identifying a clear market need, then leaving it empty for nearly a decade. Something I wish people in the US understood is that we have a responsibility to be good at our jobs, compete vigorously, do good work, and make sure that we are putting forth competitive offerings such that we don't leave giant market gaps for Chinese competitors.
Imagine clutching pearls over China when US tech companies are 100% responsible for this situation. It's not the vibe.
There was an opportunity to intervene before TikTok became a big deal (think spring/summer 2019) but of course the US missed that. Had US leadership been effective, they could have gone to YouTube and Instagram and said, "Can you please make Reels/Shorts competitive ASAP and not wait until TikTok hits critical mass?" https://www.nytimes.com/2019/03/10/style/what-is-tik-tok.html
Perhaps we could have also asked Facebook/Instagram/Twitter to make their algorithms not absolute dumpster fires of conflict and hate. TikTok is competitive because TikTok made a competitive product, and I'd like to see more emphasis on making something better than TikTok (surely if you believe in western values you agree it is possible) rather than trying to have the US government pick winners and losers. Doomscrolling is not a thing on TikTok.
P.S. Matt's take on TikTok showing Ukraine disinformation is also bad; TikTok has shown me plenty of content that aligns with the pro-western view that Russia is committing atrocities and the Ukrainians are fighting in righteous self-defense of their sovereignty. I think Twitter is equally concerning here. https://mythoslabs.org/wp-content/uploads/2022/03/Part-III-Analyzing-Twitter-Disinformation-and-Propaganda-Related-to-Russian-Aggression-Against-Ukraine.pdf
Today's comments:
10% "The damn paperboy is late again"
40% "Matt, you're crazy, this is totally unconstitutional"
40% "Matt, you're a crazy genius, just change the constitution"
10% "I saw the word 'teacher' and the word 'gender' and now I must vent my spleen unto the electro-ether lest I Hulk-out and ruin yet another outfit."
I don’t know if anyone else has mentioned this, but it seems the whole problem with TikTok is that it is a fabulous video platform which became wildly successful while video platforms created by American firms like Facebook have been utter shite. TikTok apparently excels at ease-of-use while American attempts at video apps were clumsy and lacking in features. What happened to American technological ingenuity?
Matt’s argument here (which I’ve seen in others like Ben Thompson) is absurd and hypocritical. Even if TikTok's algorithm is 100% controlled by the PRC government, it competes freely and fairly in the open US market. It’s not a monopoly and has not abused its power in anti-competitive behaviors. Banning TikTok would be philosophically contradictory to not only our principles of free speech, but also that of market economy.
Matt’s argument is that we need to bubble wrap US citizens in US propaganda. Kevin Drum takes the next logical step and says that it needs to be center left bubble wrap and Fox News is the enemy. 1984 or a Brave New World or the animal farm book, one of those, had to warn that this is sort of a tricky subject.
I have a wholly different take, which is that we need to be educating our people from an early age to recognize propaganda and influence campaigns and to harden themselves from pernicious influence, to seek facts separate from opinion, to discern data from anecdote, to think critically. We can not coddle, we can only harden.
I can't shake the feeling that this is like a deeply problematic standard. Like everything about the specifics of the case seem right, but like we seize your firm based on very little actual evidence other than well they're Chinese and probably are doing something bad.
They made a fundamentally excellent product that fills a completely different niche than home grown socials and now it's going to be taken from them. It doesn't seem like fair play even if everything about the case is correct. What firms could you not destroy under the evidentiary standard of we think they’re probably doing something bad?
I'm sympathetic to this but China straight up bans American companies like Google and Facebook and to not do something about TikTok seems like unilateral disarmament.
I believe a lot of TikTok is now run from the US anyway and we could do something like have the US audit the algorithm, but that has it's own problems and may not even technically be possible.
So make an accusation you can prove in public.
I mean we know TT is doing all these things like we knew Iraq had a secret wmd program the inspectors couldn’t find. Like what could you not take under this kind of film flam everyone thinks it but no one has proved it standard?
That does not seem like a generalized principle of law but an invitation to seize anything and everything inconvenient to our rulers.
Part of what helped us resist the temptation to seize stuff using the DPA is that when it Truman tried, the SC shot him down in Youngstown.
The Soviet Union and PRC banned all sorts of American media during the Cold War, but that didn't stop Americans from having a First Amendment right to read media from those countries. See Lamont v. Postmaster General, 381 U.S. 301 (1965), summary here: https://www.mtsu.edu/first-amendment/article/848/lamont-v-postmaster-general
To repeat myself from 11 hours ago:
If the objection is that TikTok is more effective at influencing Americans' attitudes and beliefs than Pravda was, then you've basically just invented the free speech version of the Emma Goldman quote about how "[i]f voting changed anything, it would be made illegal."
“Reciprocity” covers a range of sins, unfortunately.
And it’s not like we’re doing anything to Chinese firms that China’s government isn't already doing to them in greater measure.
So because Chinese consumers are forced by their government to use lower quality services we should accept that for Americans on a just trust us basis?
I’m quite certain that if you extirpated any of these firms (Facebook, TikTok, YouTube) from the face of the earth tomorrow, any firm with sufficient server/AWS capacity could have a near-perfect replacement up and running within the year.
As for “just trust us”, we should we looking for a Cold War precedent sufficiently narrow to be reasonably confident it won’t be abused and then watching its use closely.
If it’s so easy to replace why is Reels so inferior?
Network effects. Period.
Kill the original and they’ll reestablish somewhere else, almost instantly.
Come now, you know this perfectly well.
The only reply you’ll ever get out of me again is:
LOL.
I'm not sure what is to be done about it, but TikTok's algo is *scary* good at personalizing and maximizing engagement. Use it for a week and then think about it for a minute and it's basically the social media tech equivalent of those fantastical GPT3 word models or DALL-E pictures out of OpenAI. The tech they've made is genuinely disruptive, and their growth reflects that. Compound on that that we have little idea what levers the CCP is pushing on the backend, and purely from a consequentialist perspective, I'm pretty convinced that it should have been banned under pretty much whatever spurious pretext the State department lawyers could have ginned up.
I've used it off and on for two years, and I don't find it find particularly good at personalizing, or particularly addictive. It's moderately entertaining, and I look at it once a month or two. I can see how it sort of personalizes over time, but honestly I don't see anything incredible or amazing there- I'd argue Youtube personalizes much better
lol, the main thing it does is just not factor the creator's past popularity or follower count into relevance scores so it's not as easily game-able by bots. Also, less ideal for "high importance" creators. You just aren't used to that.
Saying "we shouldn't let TikTok provide service to Americans" is equivalent to saying "Americans shouldn't be permitted to use TikTok". It's always worth considering that censorship/deplatforming/restricting access/etc. is imposing restrictions not just on the speaker/service, but the listener/user as well.
Obviously TikTok is doing censorship itself, and the idea here is to use the threat of banning it as leverage to get them to stop/sell to an American owner that maybe won't censor users, but cutting off access to a censored service is also a form of censorship.