Discover more from The Line
Adam Zivo: Musk, Ukraine, and the rights of social media companies
The Twitter owner's apparent decision to suppress pro-Ukraine content is bringing renewed attention to the "publisher or platform" debate
By: Adam Zivo
After Twitter made its algorithm open source two weeks ago, a group of analysts discovered that content about Ukraine is being marked as “unsafe” and suppressed. This seemingly confirms months of suspicion that owner Elon Musk is warping the platform in Moscow’s favour, and, if indeed true, suggests that stronger regulations are needed to protect social media platforms from oligarchic political manipulation
Shortly after Musk took control of Twitter last October, pro-Ukrainian users began sharing anecdotal evidence of censorship. Several users with large followings claimed that engagement with their posts had suddenly cratered, and, mirroring this, smaller accounts claimed that pro-Ukrainian content was disappearing from their newsfeeds, with pro-Russian tweets inexplicably showing up instead. In some instances, popular pro-Ukrainian users, including a soldier known for sharing his insights from the frontline, were suspended from the platform for unclear reasons.
Having noticed a reduction of pro-Ukrainian content on my own newsfeed, I posted a tweet late last November asking other users to share their stories. Many did. A handful of widely-followed pro-Ukrainian users individually provided me with their account analytics, which confirmed that they had experienced sudden, severe drops in engagement and follower growth. Crucially, their dips began around the same time, just after Musk’s takeover, and were similarly sized, with engagement consistently dropping by approximately half. Something was clearly up. The only real question was whether this was deliberate or part of wider dysfunction as part of Musk’s post-takeover layoffs.
While users’ qualitative (anecdotal) perceptions of their newsfeeds are helpful, they’re also vulnerable to confirmation bias — some people could falsely believe that there’s less pro-Ukraine content on their feeds simply because they’ve been told to expect such censorship. Further complicating things, while many users emphatically said that pro-Ukrainian content was disappearing, other users saw no such change.
This made it impossible to pitch a story about Twitter censorship to mainstream publications — everything was just too speculative. However, these months of uncertainty finally ended when Twitter made its algorithm open source.
A group of analysts, led by tech expert Aakash Gupta, swiftly reviewed Twitter’s code and, though they weren’t specifically looking to validate the Ukrainian censorship theory (and even seemed unaware of it), they accidentally showed that, yes, Twitter is suppressing commentary on Ukraine.
According to Twitter’s algorithm, tweets are deboosted if they are labelled as violating Twitter’s “space safety.” Based on a screenshot of code published by Gupta, there appears to be 15 such labels related to this mechanism, covering broad categories of infractions.
Many of these infractions are unsurprising — pornographic content, medical disinformation, generic misinformation, violence, hate, and links which direct back to untrustworthy websites. However, there was an outlier: “UkraineCrisisTopic.”
“Misinformation is highly down-ranked. Anything that is categorized as misinformation gets the rug pulled out from under it. Surprisingly, so are posts about Ukraine,” tweeted Gupta in a thread explaining Twitter’s algorithm.
His mention of Ukraine was brief, as though he had stumbled upon a curiosity, and yet the revelation quickly went viral among Ukrainians and their supporters who felt that they finally had concrete evidence that their views were being suppressed.
A note of caution: it’s unclear what that label exactly means. For example, it’s possible that the label applies only to tweets that include Ukraine-related disinformation, not all Ukrainian content.
However, this would be odd as, based on Gupta’s report, Twitter already effectively handles misinformation through three broad “space safety” labels (“general misinformation,” “medical misinformation,” and “civic integrity misinformation”). COVID-19 didn’t get a specific label. Why would the war?
It should also be noted that, based on anecdotal evidence from Twitter users, misinformation about Ukraine has not actually decreased — rather, pro-Russia misinformation seems to be thriving while pro-Ukrainian content is being deboosted.
It’s therefore reasonable to believe that there’s something pernicious about the “UkraineCrisisTopic” label. This belief could be wrong, of course, but Twitter has been unable to provide clarification on this matter as, under Musk, the company has stopped replying to press inquiries (since March, Twitter’s press inbox automatically replies with a poop emoji).
It would be highly concerning but not all that surprising if such censorship is occurring, especially because Musk has a history of being chummy with Moscow.
In early October, Musk tweeted a “peace proposal” that was widely criticized for being wildly impractical and highly favourable to Russia. Afterwards, political analyst Ian Bremmer claimed that Musk had told him that he had met with Russian President Vladimir Putin before sharing the “peace plan,” a charge which Musk vehemently denies.
More recently, Musk claimed that Ukraine’s 2014 pro-western Euromaidan revolution was “undeniably a coup.” This narrative has been integral to Moscow’s information war against Kyiv and has been condemned by an overwhelming majority of experts on Ukrainian politics.
Despite positioning himself as a defender of free speech, Musk himself has a penchant for information manipulation. During the “Twitter Files” scandal last December, he publicly shared several internal Twitter documents and communications that predated his ownership of the company, allegedly because he wanted to defend free thought and expose Twitter’s history of censorship.
However, these files were provided to only a small number of journalists who Musk ideologically aligned with. Musk’s hand-picked journalists were then accused of misrepresenting the Twitter files and exaggerating their salaciousness. Critics quickly pointed out that, by withholding information from competing analysts, Musk’s commitment to “transparency” and free thought seemed disingenuous.
In December, Musk temporarily banned prominent journalists who criticized his management of the platform. Two months later, frustrated with his own declining popularity, he allegedly “urgently” ordered Twitter’s engineers to reconfigure the company’s algorithms to boost his own tweets’ rankings by a factor of 1,000, ensuring that tens of millions of extra users were forced to see his content. In the last few days, Twitter seems to have limited the ability to people to post and reply to links to articles published on the Substack (like this one), though that seems to have been at least partially reversed after sharp criticism.
It seems intuitive that an American oligarch shouldn’t be permitted to adjust the west’s infospace to the benefit of foreign adversaries — but the ethics are more complicated than they first appear.
You can break down the issue into two parts: (1) do oligarchs have the right to use their property to influence public discourse; and (2) can they exercise this influence in opposition to American foreign policy?
In the United States, where property rights are near sacrosanct, it seems that oligarchs do indeed have the first right, which, in some cases, is exercised through private ownership of media outlets. For example, Musk and his defenders have deflected criticism by pointing out that Jeff Bezos, another billionaire oligarch, received far less criticism when he acquired the Washington Post.
As for the second question, the right to free speech obviously includes the right to criticize the government and, by extension, to advocate for alternative foreign policy views. This freedom can only be meaningfully exercised if people also have the freedom to use their property in support of political speech.
However, once you get into the regulatory debates surrounding social media platforms, these arguments above, which excuse Musk, fall apart.
For around a decade, policymakers have been grappling with whether social media companies count as “publishers” or “platforms.”
In the United States, publishers are responsible for the content they share and are liable for harmful material that is pornographic, defamatory, or inaccurate. Regulators who want social media companies to play a stronger role in content moderation have, naturally, advocated that they count as publishers.
In contrast, social media companies argue that they are “platforms.” From this perspective, companies like Twitter more closely resemble internet service providers, in the sense that they’re neutral conduits of other people’s data, which they bear no responsibility for.
It’s typically argued that if you want to be a platform, you need to be politically neutral. This argument has been popular among American Republicans who have been frustrated by perceived left-wing bias on major social media platforms (i.e. rebuking allegedly biased fact checkers).
If Twitter is censoring Ukrainian content, then it is making editorial decisions and should be treated as a publisher, which, crucially, includes accountability for disinformation. Musk can’t have his cake and eat it, too.
Alternatively, some have argued that social media platforms should be treated as a public utility. These platforms are inherently monopolistic — their success requires a large, critical mass of users, which encourages market concentration at levels unseen in traditional media. At the same time, they facilitate a wide variety of communications that, collectively, are essential to society.
Monopolies (or quasi-monopolies) that provide essential services have historically been subject to tight regulation. This is typically justified on economic grounds (i.e. preventing price gouging), but regulations have also been imposed in the interest of fairness.
If Twitter is an essential monopoly, then it ought to be regulated accordingly — which means that political neutrality should be imposed upon it to the greatest extent possible. Under this framework, if Twitter cannot provide fair services, then the natural response would be to nationalize it — as has previously been done with utility and railway networks.
Until recently, this problem has been partially hidden by the fact that, until Musk’s Twitter takeover, major social media companies have been owned by large collections of shareholders, which imposes some degree of rationality on their behaviour. Corporate leaders are constrained by a fiduciary duty to their shareholders, who are too fragmented to think beyond profit-maximization.
Musk-era Twitter is the world’s first example of a mature social media network operating under the erratic, somewhat irrational leadership of one man.
Given an underdeveloped regulatory environment, there’s no real legal recourse against Musk’s (alleged) censorship of Ukrainian content. Twitter, as a private company, enjoys free speech rights, which includes the right to set its own rules on speech, however politicized or unfair, within the confines of the platform.
The only way to change that would be to establish countervailing regulations. The U.S. Supreme Court is considering hearing two cases which have the potential to resolve these issues (both involve laws barring online platforms from taking down certain online content), but, even if these cases are heard, they likely won’t reach the court until June at the earliest, or perhaps even as late as 2024. After that, it may take several more months to make a decision.
In the absence of a legal avenue, the alternative solution is to apply economic pressure — essentially, to call public attention to Musk’s alleged censorship and pressure advertisers — those few he has left.
Pro-Ukrainian voices could also push for legal action in other countries where the regulatory environment might be more responsive to policing Twitter’s biases. This could solve the problem in a piecemeal way, and would best be focussed on countries where political support for Ukraine is most consequential and fragile.
For now, though, it’s not clear that anything will change, at least anytime soon. It’s up to the consumers to be informed about what information they’re seeing on Twitter, what information they’re not, and why.
The Line is Canada’s last, best hope for irreverent commentary. We reject bullshit. We love lively writing. Please consider supporting us by subscribing. Follow us on Twitter @the_lineca. Fight with us on Facebook. Pitch us something: email@example.com