If you back serious free speech, you quickly learn three things. First, once speech regulation is suggested to government, it operates under Parkinson’s law — invariably expanding to fill the space available. The second is that, with a few honourable exceptions, inside almost every politician there is an authoritarian censor struggling to get out. The third is that organisations that might be expected to oppose censorship will often do no such thing.

The government’s Online Safety Bill, which had its second reading in the Commons last Wednesday, demonstrates all of the above. It originated in a well-meaning but incautious “duty of care” campaign instituted by the Telegraph following a high-profile child suicide in 2017. It was aimed specifically at preventing social media from targeting children with inappropriate material on such things as eating disorders and self-harm. Unfortunately, since its government takeover it has morphed into a legislative monster. It now would impose highly intrusive controls on what can be seen online by children and adults alike, on social media such as Facebook and Instagram, on facilities such as Youtube or Instagram, and in certain cases on search engines.

Under the Bill as introduced, all social media sites in the world with a substantial UK readership will indeed have to take steps, by age verification or otherwise, to protect children from a wide variety of unsuitable content. But they will have to go a great deal further.

Machinery will be required to flag and remove posts alleged to contain illegal matter: something that sounds like sweet reasonableness itself, until you realise that illegality extends to pure speech crimes such as harassment. It also covers offences under the Public Order Act, ranging from verbal attacks on religion or sexuality to abusive words likely to cause distress.

The effects of this exercise in legislative overkill are predictable

In the case of the largest sites — Facebook, Twitter and the like — there is now a new and highly controversial “legal but harmful” duty. Subject to a few protections for news providers and the like, sites will have to take steps to remove, or to limit access to, a wide swathe of material otherwise entirely lawful. Taboo content will be determined by government regulation (via a list that can at any time expand at the whim of a minister), if deemed “harmful to adults”. Content will be censored if it “presents a material risk of significant harm to an appreciable number of adults in the United Kingdom”, even if that harm is not specified.

Any breach of these duties is liable to be met with a very sizeable fine from Ofcom. In the last resort, a court order may block access.

The effects of this exercise in legislative overkill are predictable. Although sites must have a procedure allowing users and others to appeal takedowns or restrictions, and also face a vague duty to “have regard to the importance of protecting users’ right to freedom of expression within the law”, it will undoubtedly lead to a “censor now, argue later” culture.

Faced with allegations from an active pressure-group representing, say, trans, LGBTQI or fundamentalist religious interests, that a post amounts to offences against the Public Order Act, sites are likely to take the safe way out. Aiming to avoid a possible spat with Ofcom, they will either take down the posts complained of, or drastically restrict access to them.

The scope this bill would give to future governments for ad hoc censorship is frightening. Had the Bill been in force during Covid, it would have been perfectly possible for an administration, at the stroke of a minister’s pen, to require Facebook, Twitter and the like to remove, restrict or censor anti-lockdown or anti-vaccination articles. Ministers need only specify such content as likely in their view to cause serious harm.

Even if no such order were made, Ofcom could well threaten to take the view that such matters were harmful anyway and thereby obtain their removal or restriction. (If this seems far-fetched, it intervened fairly heavy-handedly in 2020 when the broadcast media, over which it already has control, had the temerity to produce material casting doubt on the official position.)

This is merely the Bill as drafted. Although a few MPs were refreshingly alive to the threats it posed to free speech (stand up John Whittingdale and Adam Afriye from the Government benches, and Darren Jones from Labour), the comments on it during the second reading show a clear and present danger of more mission creep to come.

There was, for example, a litany of calls for the “legal but harmful” provisions to apply to all sites, including minnows such as BitChute, Gab and Minds. This shows that an appreciable number of MPs took the Bill as a call not simply to rein in the worst excesses of Silicon Valley, but to sanitise the net from carrying material they did not like.

There were similar calls — including from Opposition spokesperson Lucy Powell — for the Bill to cover misinformation generally, thus giving Ofcom a kind of roving commission to protect us as far as possible from controversy and propaganda deemed to be bad for us. Other individuals called for specific restrictive provisions about matters such as misogyny, incelism, fraud, self-harm and the like.

The abrogation of the right to anonymity, except in carefully-defined cases, was seriously advocated. Margaret Hodge, for example, argued — apparently quite seriously — that all users should face identity checks from a “third party with experience in unique identification” so that the name and address of every single user would be available to law enforcement.

The ball is now in the court of free speech supporters

The Bill now goes to committee. We do not know how many of the above changes will be pressed, or simply accepted by the government as the price of a quiet life. There is now a distinct danger that this legislation will balloon further so as to become not simply a corralling of Big Tech, but a general censorship measure. It may morph into a UK Great Firewall aimed at seriously curtailing for our own good what we (or at least those of us not tech-minded enough to equip ourselves with a VPN suggesting we are outside Britain) can see or watch online.

Silicon Valley and Big Tech won’t seriously complain. They already have the machinery to control what appears and what doesn’t, and to kick people off that they don’t like. Whatever Elon Musk may say following his acquisition of Twitter, it’s not in its nature or its business model to be concerned with free speech as a matter of principle. If anything it may even welcome attempts like this to regulate speech at the state level, if only it can take the expenses of regulation in its stride better than its smaller competition can.

The ball is now firmly in the court of free speech supporters, libertarians and organisations like the FSU to make a fuss. All must publish widely the dangers faced by the free internet from both the original Bill and the proposals to extend it. They must use all possible means to get their views across to MPs during the committee stage. If they don’t, the UK is likely to become not so much the “safest place in the world to be online” (the government’s slightly chilling phrase), but rather the most boring and also, more seriously, one of the least informative.