In recent weeks the British people have come to know and fear the phrase "circuit breaker". For most of us, it means a short-term lockdown that is suddenly imposed when coronavirus infections start to spiral out of control.
On Wednesday, though, a different kind of circuit breaker – less well-known but perhaps just as consequential – catapulted Silicon Valley headlong into a new level of conflict with President Donald Trump and his allies, with free speech, national security and possibly the outcome of the US 2020 election at stake.
That morning the New York Post, a redoubtable conservative tabloid controlled by Rupert Murdoch, published a front page scoop containing what it said were leaked emails from the Ukrainian energy company Burisma to Hunter Biden, the troublesome son of Mr Trump’s opponent Joe Biden.
Within hours, Facebook had "reduced" the story on its service, instructing its algorithms to rank it much lower in users’ needs feeds until independent fact-checkers could have their say.
Twitter went much further, blocking its users from tweeting the link or sending it to each other by private message, as well as telling users who clicked on existing links that the story might be "unsafe". It even locked the account of White House press secretary Kayleigh McEnany.
Viral circuit breakers
Why? Both companies must have known that the decision would prompt fury. Their motivation lay in the years of criticism they have endured from every direction all over the world for the ways their networks have systematically boosted sensational and false information.
To stop that from happening, tech critics have sometimes mooted the concept of a viral circuit breaker. The idea is that social networks could build systems to notice when a potentially dubious story is going viral and temporarily halt or slow its spread while it can be examined more closely.
Facebook already does the first part of that, using artificial intelligence (AI) to spot likely candidates for its outside fact-checking programme. Until recently, it only suppressed those stories after they had been rated false – a prudent but long-winded process that meant flagrant nonsense often reached millions of people before being throttled back.
A few months ago Facebook started to "reduce" articles preemptively if certain "signals" suggest they are deceptive. It was a new practice, so most people had not heard of it. It certainly had not yet been used against a major mainstream media institution.
Now it has been. From the start, some researchers and journalists questioned the authenticity of the emails, which had a complex backstory involving Mr Trump’s former lawyer Rudy Giuliani and had come at just the right moment for a fabled "October surprise". A key figure in the story also seemed to be confused about his role.
Then Twitter made an explosive claim: that the New York Post’s article contained, or might contain, what it considered to be "hacked materials". Facebook made no such claim, but reporting from the New York Times suggests it was also alarmed about the same possibility.
A repeat of 2016
The claim was an obvious flashback Russia’s highly successful intervention in the 2016 US election, in which documents stolen from the Democratic National Committee and the Clinton Campaign inspired numerous negative news stories that dominated media attention for weeks at a time.
Kathleen Hall Jamieson, a non-partisan sociologist who has been studying American elections for decades, has argued that those leaks were Russia’s most consequential intervention, probably enough to swing the vote.
Russia tried the same thing in France in 2017, adding its own fakes to a cache of emails liberated from President Emmanuel Macron. Experts call this a "tainted leak": concocted evidence shrouded in a cloud of genuine material.
In Britain, officials believe that leaked Brexit negotiation documents brandished on the 2019 campaign trail by Jeremy Corbyn also ultimately came from a Russian hack. At the time, Facebook’s former chief security officer Alex Stamos told the Telegraph that Mr Corbyn’s decision had been "incredibly dangerous".
The same hacking units are clearly still operating today. In January, Russian hackers were caught breaking into Burisma itself, while Microsoft says it has seen three separate hacking campaigns against both parties by Russia, China and Iran (it must be getting crowded in those servers).
To cap it off, the New York Times now claims that, behind the scenes, US spooks have been "picking up chatter" (that is, intercepting signals) indicating that Russia is preparing an October surprise using the stolen Burisma emails, perhaps woven together with forgeries. Facebook, too, warned about this last month.
The arbiters of truth
So what? Even if there are good reasons to question the Post’s story, why should gigantic social media companies ask such questions for us? Who made them the judge of what is true, over the heads of politicians, governments, researchers and the press?
Many argue that the problem is that both companies are, to a great extent, doing exactly what has been asked of them. Stung by evidence that their algorithms are biased towards sensationalism, partisan conflict and even ethnic cleansing, they have heeded worldwide demands to stop misinformation from running rampant.
Breaking up big tech | Who wants to curb the power of FAANG?
These demands come from the uncomfortable truth that Facebook and Twitter already routinely "censor" the world in their own commercial interests. They presume to tell users what we should see at the top of our news feeds in order to make maximum profit from our attention. They have always removed content that would ruffle their advertisers or cause legal risks, even if we want to see it.
For that reason, the spread of fake news is not just some organic problem to be blamed on human nature, but a consequence of the commercially-driven censorship that social networks already undertake. If they already have their thumbs on the scale, why shouldn’t they heed other concerns?
Censor with gusto
In 2018, this dynamic looked like a ratchet that seemed to only going in one direction. Each time a social network intervenes in politics, it strengthens the precedent that it will do so in future. That incentivises partisans to work the referees, slowly depriving companies of excuses to stay aloof.
Today the coronavirus pandemic has turned that ratchet further and faster than many ever imagined. Facebook and Twitters’ early decision to censor coronavirus misinformation because of its very real potential to cause harm has opened the floodgates to an unprecedented level of enforcement against even the President of the United States.
Yet the companies could easily have interpreted their rules this way several years ago. There are stacks of past examples where they appear to have gone easy on politicians or media outlets because they did not want to ruffle feathers. Perhaps they are finally thinking like Shakespeare’s Richard III: since they will never escape accusations of censorship, they might as well censor with gusto.
If that is surprising, it is only because our society has not yet fully reckoned with the power that these companies have, for good or for ill. This week they forced us to, and the real shock is that they didn’t do it sooner.
Свежие комментарии