Toward an Uncensored Internet

Most social media platforms occupy an uneasy middle between these two fundamental modes. They have some content standards, but not many, nor particularly rigorously enforced. The companies rely upon, and empower, a tangle of decentralized content creators, but host and govern them using centralized infrastructure that is vulnerable to influence operations.

Facebook told the Senate Judiciary that it plans to employ 20,000 or more moderators by the end of 2018, which will be roughly one moderator per 70,000 daily active users; one moderator per 106,500 monthly active users. And this is after political pressure provoked a hiring push. In such an environment, consistency is not just difficult to achieve, but frankly impossible.

The difficulty compounds when the cultural differences across user populations come into play. Neither can the moderators themselves be expected to adopt a single perspective, especially since many of them are hired in lower-​cost countries than the United States. They are given seconds to make each decision.

Social media companies’ tendency to selectively cave to users’ demands doesn’t help. It is natural for a company to heed its customers, but Silicon Valley behemoths are most attuned to the vocal users who are their friends and colleagues. Famed investor Marc Andreessen called the industry “extremely left-​wing,” pointing out how overwhelmingly the region’s voters favored Hillary Clinton.

Research from the Stanford Graduate School of Business demonstrated that “technology entrepreneurs support liberal redistributive, social, and globalistic policies” (albeit “conservative regulatory policies”). The Lincoln Network concluded, based on its own survey, that tech employers’ political homogeneity is “alienating a significant talent pool.”

Twitter’s verification program is a prime example of the tense balance that social media companies struggle to maintain. Originally it was intended to do something simple: Verify that a person was who they said they were. Verification was meant to quash celebrity impersonators and other deceptive accounts.

Celebrities have long been a crucial constituency among Twitter’s larger user base — so much so that their presence qualifies for a dedicated Wikipedia page! — and verification protected them from impersonation. Twitter was also incentivized to keep the brands using its platform happy, in order to maintain the flow of advertising money. Nobody appreciates a program to devalue copycats like the trademark enforcement team of a large consumer company.

Because Twitter granted the vaunted blue check only to public figures, it automatically became a sign of status. That was probably unavoidable in the beginning. But then, instead of verifying anyone who demonstrated their identity, Twitter veered down a different path. The company doubled down on treating verification as prestigious, developing special features to insulate the verified from the plebs, and even launching a separate app for popular and verified users.

Before long the rest of the user base perceived the blue check as a mark of endorsement by Twitter the company, not merely a validation of accurate identity. The problem was that Twitter had never given up on the original meaning of verification, so it kept on verifying figures reviled by many of its users. Given the mixed messages the company had sent about verification, users interpreted the blue checkmark bestowed on a person like white supremacist Richard Spencer as a gift of status directly from the company.

The progressive core of Silicon Valley was particularly enraged, as this perceived offense was piled on top of the long-​mounting fury about hate speech and abuse on Twitter. Twitter underwent a sort of crisis of faith as criticism reached this latest high, and de-​verified prominent white supremacists.

The company left itself in an awkward position. Now Twitter will be considered responsible for its verified users’ off-​platform behavior, should their antics meet an ambiguous and ever-​shifting standard of odiousness. On top of that, Twitter will be considered morally culpable for the offensive-​to-​at-​least-​one-​group remarks that verified users cannot be relied upon not to make.

Here’s the end result: Richard Spencer is easier to convincingly impersonate. Imagine if someone tweeted from a spoof account that a redux of the Charlottesville killing was imminent. Can you imagine the ensuing panic? It hasn’t happened, but it’s not beyond plausibility. Surely this risk is not consistent with Twitter’s overall efforts to reduce conflict and misinformation on its platform. A verification program that allowed any user to opt in and validate their identity would reverse the problem.

If Twitter can send users vague notes alerting users that they supposedly interacted with Russian propaganda, surely it can sort out the contradictions in its verification program. Eventually. But what happens when the group of users who can prompt Twitter to make reforms, at long last, are politically homogeneous?

Twitter is a private company, so of course it can listen to whoever it wants. The website is a big player in the market for real-​time news commentary, which may seem like an insignificant niche until you consider that politics, tech, and finance heavily rely on the platform to understand the world — not to mention media, which then amplifies those paradigms.

Twitter is not the only company making users who prioritize free expression uneasy. Verification is just one example of the opaque and unaccountable systems that govern public discourse. Over the course of 2017, America has been wrestling with how to address the growing power of Facebook, Google, and less well-​known Silicon Valley peers. The almost total lack of state oversight has become apparent to both sides of the political aisle, leading to calls for regulation.

In October, 2017, New Jersey congressman Frank Pallone issued an open letter to the CEOs of Google, Facebook, and Twitter, requesting that they explain the rationale behind their content policies. “With a goal of ad clicks or driving page views, these companies’ policies are not neutral; they actively shape content on the web,” Pallone wrote. “And to the extent that these companies’ platforms have publicly available policies for moderating content, those policies are vague and applied inconsistently. This lack of transparency makes it difficult for consumers to understand how content is controlled and for the government to oversee the market.”

The internet is often called a democratizing force, but its tendency to enable natural monopolies more closely resembles feudalism. In fact, they thrive based on the free content that users constantly create and upload. Facebook would be nothing without status updates, and Google would be nothing without countless websites to index.

Huge tech platforms are the internet equivalent of nation-​states, and it’s increasingly common to use this metaphor to discuss them. Facebook and Google in particular, with their dominance of information flows and quasi-​duopoly on advertising, act like empires in which regular people can only be serfs.

Public figures with a sufficiently large and passionate contingent of fans, like the upper echelon of YouTubers, are akin to aristocrats with large estates full of serfs — platforms are careful to avoid their defection, because it would be particularly costly. The most engaged users are a great source of ad revenue. But if you’re like PewDiePie and your transgression is verboten enough among the elite chattering classes, expect to be penalized.

When regular users run afoul of the algorithm, or are dinged by a moderator, it’s common for them reach out to accounts with larger followings. The hope is that their pleas will be sent to the top by a sympathetic intermediary. It’s not unlike begging a duke to bring your grievance to the king. Jack Dorsey doesn’t care if some random Twitter account gets shut down, but he might lift a finger if a sufficiently prestigious bluecheck brought it to his attention. Or at least that’s the theory.

Neither Facebook nor Google can decide what people say, but their respective policies and algorithms decide which kinds of speech will be surfaced to the broader public. Those policies are largely driven by the profit motive, often based on the respectability standards of advertisers, as New Jersey’s Pallone noted. Sufficiently broad public outrage can prompt specific reforms, as with Twitter’s bout of de-​verification, but relinquishing power is not on the agenda. Mark Zuckerberg doesn’t need to run for president; he is the king of 1.32 billion users who live in his realm every day, all of whom are mandated to use their “real” names.

It is true that users can emigrate, so to speak. They can leave a given platform — but where will they go? The internet doesn’t have limited territory, since anyone can put up a website or run an email server. Rather, the scarce resource online is attention; access to a potential audience.

Internet idealists talk about how freedom of information opens up opportunity for everyone. You can learn whatever you want, build whatever you want, and communicate whatever you want. For example, instead of needing thousands of dollars to self-​publish a physical book and buy ads for it in a magazine, you can make an ebook for free, distribute it however you like, and promote it on social media. Compared to the previous status quo, this is a genuine improvement! The internet idealists are right: Opportunity truly is more broadly available than it used to be.

But what if Amazon bans your ebook? What if Barnes & Noble and Kobo also deem it inappropriate? The unspoken catch of the internet as democratizing force is that if you are weird enough, or aberrant enough, and you either use the wrong keyword or attract the baleful eye of the administrators, you’ll be banished. In that case, it doesn’t matter that there are fewer gatekeepers — the handful of big, flourishing gatekeepers are key, and they have shut you out.

It may be that there is another path. Disenfranchised hackers and crypto-​anarchists are building parallel institutions that no one group can own or control, instead of trying to force giant tech platforms to accommodate them. They recognize that any entity that prizes advertising revenue above all else can’t be relied upon for civic neutrality.

The point is not to build a safe space for hate speech (well, most of the time). Rather, the cypherpunks recognize that dissatisfaction from Unilever and its ilk can snowball into a YouTube demonetization epidemic that disproportionately limits the earning ability of smaller creators; that LGBT commentators will be collateral damage in the pursuit of a friendly environment for advertising. And because it’s all automated, users will have no recourse.

The answer is decentralization: Spreading out power to numerous participants in a network or platform. People have talked about decentralization since the earliest days of the internet, sometimes seeing it as the key grand principle behind the whole project. Currently we’re witnessing a renaissance of decentralization, led by a handful of prominent projects.

For example, bitcoin and the proliferation of new cryptocurrencies have taken off. Bitcoin doesn’t derive its value from the backing of a government, nor from a corporate overlord with control of its features. Bitcoin’s value stems from its technical design, and because the code is open-​source, anyone can fork away from policies they don’t like. The asset is “hosted” by people all over the world who have nothing to do with one another.

As another example, the censorship-​defying protocol Tor anonymizes browser traffic and provides a way to publish websites that no commercial hosting company will allow. The power to dictate terms can’t accrue to one chokepoint. An onion service doesn’t rely on the persistent goodwill of a domain registrar; instead it relies on cryptography.

The downside of decentralized technologies is that they are rarely user-​friendly, and often require a greater tech savvy than the general public possesses. Using a federated social-​networking protocol like Mastodon, which mimics Twitter but diverges in a number of confusing ways, requires a higher level of mental investment for a lower payoff (since its network effects are weaker). And the most common response to Urbit, a project intended to radically decentralize web-​hosting, is bafflement.

On the bright side, it currently looks like money is the “killer app” of decentralization. The potential to profit, or be able to transact in otherwise circumscribed ways, is a compelling reason to read a how-​to blog post, even for normal people. Bitcoin, ether, and their proliferation of siblings may undergo a devastating crash in the next few months, or the next year — but the prices will rise again eventually.

Traditional payment processors dealing in USD have an ongoing habit of acting as a chokepoint. Hatreon, a crowdfunding site that would let anyone join as long as they didn’t use the platform to break the law, is no longer able to take pledges because their payment processor dropped them, and so far competitors haven’t been eager to take its place. Starting in 2013, the Obama administration pushed an initiative that was literally called Operation Choke Point, which put pressure on financial institutions to cut off controversial businesses’ ability to transact.

When it comes to Facebook, Google, Amazon, and their ilk, it’s unlikely that decentralized alternatives will overtake the feudal superpowers. At least not until the internet undergoes its next paradigm shift, whatever that may be. But on the fringes, alternatives do exist. A divergent path is possible. You can own your own digital territory, the borders secured by math.