Four years in the making, the Online Safety Bill has now been sent to senior ministers for review — a process that allows them to protest, to shout if anything obvious that has been missed. In this case, the process is invaluable because something huge has been missed. The Bill, if passed, would empower the Silicon Valley firms it’s designed to suborn. It would formalise and usher in a new era of censorship of UK news — run from San Francisco. This Bill would backfire in a way that its Tory advocates have so far proven unable to understand let alone address. That’s why it needs to be halted, and a rethink ordered.
The original aim is to make Facebook, YouTube, Twitter and others liable for any genuine filth promoted online. But introducing a new test of ‘legal, but harmful’ would vastly expand a bot-censorship system that is disfiguring public debate already. Politicians struggle to understand this, thinking instead that the BBC is their proper target. For all its faults, the risk of BBC bias is as nothing compared with the risk of asking woke Silicon Valley executives to police digital news by deciding between them what’s ‘harmful’ and what’s not.
I write about this in the Telegraph today. But I’d like to say more about how smaller publications like The Spectator, who rely on digital reach to grow, are already now doing daily battle against the censorship bots in a way that already sends a message from Big Tech: stay away from controversial topics, or your business will suffer.
Take Spectator TV, our new(ish) weekly show. When we had Lionel Shriver reading out her column, it was ‘demonetised’ (not given advertising money) by YouTube (owned by Google). We weren’t told why. But it seems to happen when we post content critical of Covid policy. Other times, we notice an episode has had a fraction of the traffic it normally gets. This will be because it’s been dropped from YouTube’s recommendation system (that drives 70 per cent of its views). We’re never told why: YouTube doesn’t do explanation. Usually a bot will have performed an audio-scan of the episode and found discussion of risky topics. YouTube has no instinctive desire to please the government but it wants to avoid confrontation with regulators, so can programme bots to punish opinions it disfavours.
I’d like to be frank about what we owe the Big Tech companies that I critique here. YouTube’s technology allowed us, for free, to set up Spectator TV. Social media generates a quarter to a third of our traffic on a good day: these innovators have created audiences that The Spectator has accessed and has allowed us to almost double sales. YouTube has been in touch with us to offer valuable advice on improving Spectator TV. But the team can do nothing about the censorship bots. When you are punished by the bots, you’re lucky to get an email explaining why. It’s up to you to join the dots. Criticise vaccine passports, your episode audience drops by 80 per cent. Go figure.
The business logic is clear: to maximise your audience, don’t be too troublesome. Don’t ask difficult questions about vaccine passports, or Covid policy in general. Once, we had a WHO official on Spectator TV discussing the drawbacks of lockdowns with Andrew Neil — and as he spoke YouTube slapped a health warning telling viewers to seek WHO advice. YouTube needs to censor millions of hours of content a day, so its judgements are crude. But if they are asked to censor for ‘harmful’ material, it will get a lot worse.
When we asked Professor Carl Heneghan and Tom Jefferson to review the evidence on face masks, this so displeased the Facebook bots that they labelled it ‘false information’. Why? Several times we ask Facebook to identify a word that these academics said that was false: it does not reply. No one questions the bots. The British Medical Journal spotted how sinister this trend is, how chilling for debate — like us, the BMJ will need digital traffic to keep its revenue alive. But what if we’re both told that questioning medical consensus is bad for our business?
Even worse is that Facebook can blacklist a publisher who prints what they regard as wrong-think. You’re not deleted, but your content isn’t promoted — and you’re not necessarily told why or even whether you have been selected for its sin bin. I published the face mask story on my personal Facebook feed, but not on our official Facebook page. Why? Because we get between 6 and 12 per cent of our traffic (and, ergo, sales) from Facebook: we can’t risk being on its blacklist.
As a publication, we’re too small for Facebook to answer our emails — so we can’t even ask what’s happening. We run their ads (as believers in free speech) but no, their ad agency will not ferry a message to the mothership. Silicon Valley is a force that may regulate the smallest detail of British publications but is answerable to no one in this country. The Spectator, for our part, is regulated by Ipso: if we libelled the work of an academic as ‘false information’ we’d be forced to either explain ourselves or withdraw.
The Spectator is lucky to be owned by a family, the Barclays, whose only instruction to me as editor is to be true to The Spectator’s historic mission. We’ve done all kinds of crazy, loss-making things over the years (going to the High Court over the Alex Salmond case last year for example) that cost shed loads of money but our proprietors backed us because they believe in our journalistic principles. They repaid furlough money last year out of principle. They’d rather run The Spectator at a loss than see us cower, in any way, to a digital censor. But if The Spectator was owned by a venture-capital company, on a penny-pinching remit, the financial logic is clear: don’t pursue any subjects that will cause you to be downrated, censored or blacklisted.
The Online Safety Bill would make all this far, far worse. Nadine Dorries has suggested that Jimmy Carr’s gypsy joke would fall foul of her new regime. So already, this wooly definition of ‘harmful’ expands from pornography to the edgier gags of stand-up comedians on Netflix. Just as ‘hate crime’ can be expanded to include Tony Blair insulting the Welsh on TV, ‘harm’ will be expanded by this bill to mean anything edgy. And publications will be punished, digitally, as a result.
The Spectator’s sales are strong, as I outlined yesterday. But the vast majority of our sales now come via online. So I wish I could say that falling foul of Silicon Valley does not matter to the world’s oldest weekly — but it does. The dilemma is that private companies have between them created what is now the public sphere, but with algorithms are already exerting censorship in a way that Britain has not seen for centuries. Murdoch, Hearst, Beaverbrook: none had the power that Clegg and his ilk now have over news seen by millions.
One final thing. Let’s look at what has happened already:
- The Socialist Workers Party had their Facebook page suspended. No reason was given and if it were not for uproar, it would never have been recovered.
- Novara Media had its YouTube account suspended. I’m not on the same page as Ash ‘literally a Communist’ Sarkar but think she makes political debate more vibrant and that our political space is more varied for Novara. But something they said offended the censorship bots and their page — in fact their whole modus operandi — was taken down.
- YouTube removed a video of David Davis giving a speech against Covid passports during Tory conference citing ‘medical misinformation’. But as always, it does not specify what content had breached the rules.
- @PoliticsForAlI, Twitter’s No. 1 independent feed, was deleted without any convincing explanation. A reminder about how Twitter can delete anyone, for any reason.
All of this poses huge questions for our democracy and society, but the Online Safety Bill recognises none of these problems and would instead make them much, much worse. ‘Legal but harmful’ is a phrase that would publish any publication that goes against the grain. Facebook had banned stories suggesting that Covid leaked from a Wuhan lab, seeing it as a Trumpish conspiracy. It’s vital that such avenues are explored by a free press because sometimes, the conspiracies are right.
The Tories may like the idea of Big Tech answering to the whims of a Tory government. But as ministers assess the Online Safety Bill this weekend they should ask: what if the ‘harmful’ definition was being set by Jeremy Corbyn?
That’s why any minister who gets the chance this weekend should suggest that the Bill is halted — and a proper discussion is opened about the implications of algorithms, bot-censorship and how free speech can be protected in the digital era.
Got something to add? Join the discussion and comment below.