Daniel Greenfield, a Shillman Journalism Fellow at the Freedom Center, is an investigative journalist and writer focusing on the radical Left and Islamic terrorism
300 hours of video are uploaded to YouTube every minute. The Google service’s effort to implement its new ban on Nazi content ran afoul of the error rate of trying to cope with these impossible numbers.
The casualties included history teachers who had uploaded historical videos, activists documenting Neo-Nazi rallies, and people uploading Hitler meme videos from Downfall. Despite Google’s boasts about its algorithms, they’re no more accurate than the ones that flag any comment which mentions Hitler.
Algorithms are good at making positive recommendations, but despite a cottage industry of machine learning censorship that took off after Trump’s victory and the media’s fake news moral panic, they’re not any good at filtering out negative content. Deeper analysis can spot how an item is promoted, leading to claims of inauthentic activity and bot networks that now fill the news, but cannot pass judgement on the actual content. But despite that, the dot coms will keep on trying.
YouTube is especially motivated because unlike Facebook, it isn’t profitable. In 2015, the Wall Street Journal suggested that despite a billion users and $4 billion in revenues, it wasn’t making it money. Since then, Google has become more mysterious about the numbers resulting in some wrangling with the SEC. The service’s varying efforts to solicit a paid subscription model that would compete with Netflix and Prime have failed. YouTube Red only managed 1.5 million subscribers. Last fall, Premium’s originals went public. YouTube TV managed to gain 1 million subscribers. That’s less than half of Sling TV’s numbers.
Content moderation is expensive and ugly. The content that really needs to be moderated isn’t political, it’s violent and sexual, running the gamut from videos of bloody car crashes to pedophilia. The cheapest way to do content moderation is to outsource it to a third world country where the staff won’t have to be paid as much and where their therapy bills won’t show up in Silicon Valley. But you can’t expect a content moderator in the Philippines to understand the difference between a history lesson about Hitler and a video glorifying Nazi Germany, or a video documenting a hate rally and one celebrating it.
Even gore and sexual content require cultural judgement calls. Should the infamous photo of the ‘Napalm Girl’ be censored? Facebook did in 2016. But politics requires even harder judgement calls.
And no amount of content moderation can keep up with 300 hours of video uploaded every minute.
Censorship is futile on the global scale. Anyone who genuinely wants to bypass YouTube’s restrictions will. What it actually does is deplatform and demonetize successful conservative content creators. When YouTube objected to my videos, I created a new channel. But I only have 67 subscribers and I don’t monetize. For Steven Crowder, one of YouTube’s targets, the demonetization is devastating.
That’s true not only of YouTube, but of social media censorship of conservatives in general.
Censorship can’t truly shut down conservatives, but it keeps conservative voices from having a secure foothold in social media, while offering every possible benefit and subsidy to their leftist rivals.
Any active conservative knows that he can be silenced, at any moment, by a motivated crybullying campaign backed by lefty media outlets, as Crowder was, or by the sweep of the algorithmic scythe. The overall effect is to make it clear that the internet and social media are the native territory of leftists. Conservatives can only be occasional trespassers, living on the internet as tokens or guerrillas.
There can hardly be a formula more conducive to a state of political radicalism.
Censorship won’t make social media less divisive, as its advocates insist. The echo chamber of the press, radio and television that the media, named after these three mediums, dreams of returning to is impossible. What censorship really does is radicalize activists by giving them another enemy to fight.
YouTube’s censorship campaign, like those of other social media companies, is based on secretive algorithms, black box moderation, and confusing policies that are being selectively implemented. The big dot coms and their media inciters pretend that cleaning up the internet can be easily done. In reality, they can’t even clearly state the problem in a way that would stand up to any regulatory oversight.
The new YouTube policy cracks down on "videos alleging that a group is superior in order to justify discrimination, segregation or exclusion based on qualities like age, gender, race, caste, religion, sexual orientation or veteran status." Will that policy be applied to videos like, “White People are Evil”, “Dear White People, Kill Yourself Now”, and “All White People are Racist”? The Nation of Islam’s YouTube channel is humming along with offerings like, “Critical Thinking Outlawed as Anti-Semitism”.
Taken literally, it would ban a video arguing for raising the drinking age or the driving age.
But policies like these are not meant to be taken literally, they’re something to point to so that the censorship of conservatives appears to be the result of impartial rules, rather than partisan malice.
The original media panic over fake news has been renamed as a campaign against disinformation. But what’s disinformation? After protests from Iranian media types, the State Department stopped funding the Iran Disinformation Project which had been working to challenge pro-Iran propaganda. According to the Iranian mainstream media people, the disinformation project’s personnel are trolls. What the definition of disinformation and trolling remains in the eye, the heart and the mind of the beholder.
One man’s disinformation network is another man’s advocacy campaign. One woman’s troll is another woman’s impassioned activist. Tactics, such as paid accounts and sock puppets, are used by all sides.
Information is disinformation if you disagree with it. Activism can be trolling if you don’t like its message.
The vagueness isn’t a bug, it’s a feature.
When the rules are unclear, all the standards are double standards. Dot coms hide behind the illusion of policy when, as Steven Crowder’s case revealed, their responses aren’t determined by written policies, but by pressure campaigns and gut reactions, with the policies as little more than retroactive excuses.
The failure of YouTube’s latest moderation effort has prompted calls for even more comprehensive crackdowns. Slate, whose origins go back to Microsoft, published a screed implying that anyone uploading YouTube videos ought to be subject to FCC approval and the Fairness Doctrine.
"It could mean rules against broadcasting hateful views or disinformation to large audiences," April Glaser argued at Slate. "Freedom of speech isn’t the same thing as the freedom to broadcast that speech."
But then what is freedom of speech. The right to write something and seal it away in a drawer? The freedom to whisper something in your own ear? The inalienable right to think to yourself?
What would FCC supervision of 300 hours of video uploaded a minute even look like?
The FCC has 1,688 employees. It would need more like 168,800 to even try and keep up with that river of content. And a tip line to report offensive speech to a flying squad of censors out of Oceania.
There is a more American solution that wouldn’t turn us into an obscene Orwellian cartoon.
It’s called freedom of speech.
Freedom of speech means letting people say whatever they want as long as they aren’t engaging in criminal behavior like sending death threats, molesting children or otherwise breaking the law.
The internet, and dot coms like YouTube, weren’t built through the careful curation of content by a discerning publisher, but as platforms for everyone. The platform that thought episodic adult sequels to the Karate Kid movies was a great content strategy for a premium service obviously has no clue.
Trying to transform platforms into publishers is a disaster that would kill YouTube, not to mention Facebook and other social sites that have thrived as little more than tools with algorithms.
The censorship of the internet is a futile project conducted in bad faith for bad purposes. Like the censorship projects of every totalitarian regime and ideology, it will fail. The only question is how many people will be hurt along the way. Speech can’t be stopped. People, individually, can be destroyed.
That is what is at stake here.
YouTube’s latest disaster is a painful demonstration of the limitations of moderation and censorship. But the internet was never meant to be moderated. And trying to do so is a doomed proposition. There’s too much content, too many people and machine learning can never compensate for human cunning.
Free speech isn’t just a good idea. It’s not a nice slogan. Speech wants to be free.
It takes more energy to censor than to speak. Hunting down and silencing people you don’t like is a lot more work than making your own argument. Entropy is not on the side of the speech suppressors.
Natural rights are not just a philosophical position. They’re innate human realities. Repressing them is the equivalent of damming a river. It futilely restrains a natural force that will eventually break free.
YouTube tried to dam the river. But the river, the billions of hours of video, will always overflow.