A Free Speech Bill for Britain

It is time for the UK free speech movement to stop playing defence and to go on the attack

By Preston J. Byrne, Senior Fellow, Adam Smith Institute

The Model Bill is available here as a publication of the Adam Smith Institute. The full Model Bill may be downloaded at the bottom of this post.

Early feedback from lawyers on this proposal, which has been fully anonymized, has been collected here.

Over the course of the last year, I have defended Americans from the UK’s Internet censorship agency, Ofcom – as well as parrying inbound censorship demands from Australia’s eSafety and several European regulators. Most of you reading this will have learned about this work, and these cases, because of an e-mail we sent to Ofcom in response to that agency’s failed attempt to fine my client 4chan

That e-mail contained a cartoon hamster, and an X post about it went viral two weeks ago.

The hamster thing is a running private international law joke between 4chan’s legal team and Ofcom’s legal team – although I suspect, as between us, only 4chan’s legal team finds it funny – about service of process formalities. Specifically, it illustrates that Ofcom’s e-mailed fines to my clients are not validly served in the United States.

Ofcom’s letters may, therefore, safely be shredded for my pet hamster’s enclosure without any enforceable legal repercussions. The joke even made it into court filings in our clients’ lawsuit against Ofcom, after Ofcom brought it up in their motion to dismiss and we wound up having to explain the joke to a U.S. federal judge.

Hamsters aside, of any foreign agency, Ofcom has been our most serious, and determined, opponent. For over a year, we have fought them, and every other foreign censor we encounter, with the objective of preventing them from laying a finger on any American citizen or company, using all lawful means available.

Fighting Ofcom is, however, a sideshow. It is not, and never has been, the main event. As between Ofcom and 4chan, and regardless of the outcome of the D.C. litigation, and however much Ofcom might bluster and engage in enforcement theatre, on the question of whether Ofcom can collect its fines against Americans, American law is firmly on 4chan’s side. Ofcom will never collect a single penny of its fine against my client.

Nor is my client’s fight against Ofcom and across America’s borders the fight I care about the most – although I care about that fight a great deal. The reality of it all is that the refusals of Ofcom’s orders, the lawsuits in America, the GRANITE Act foreign censorship shield bills now advancing in legislatures across the United States, the cartoon hamster – all of these are early battles in a far wider war.

The big fight, the real fight, is to restore free speech in the UK. Publishing this Model Bill, today, we mean to start it.

The Problem

The United Kingdom does not have free speech, not in any sense that an American would recognize. This is not a uniquely Starmer or Labour problem. This has been the case for decades. The only difference between the censorship of 15 years ago and today is the quantity in which it occurs, not the nature of the restrictions. 

The objective of the Model Bill is, as far as possible, to get the UK’s government out of the business of regulating opinions. The Model Bill accomplishes this by abolishing huge swathes of criminal offences relating to the causing of offence, insult, alarm, or distress, and by eliminating content-based speech restrictions which threaten a speaker with prison for little else than having the wrong ideas and daring to say them.  

In a free society, fools, bigots, and assholes get to speak and remain free men. That is not the price of liberty. It is liberty, and the rest of us get it too. The Model Bill’s authors are aware that this Model Bill, if enacted, would decriminalise expression that we find morally repugnant.

We accept and embrace that consequence. But we ask the reader to consider who these laws actually catch. Overwhelmingly, victims of the UK’s censorship state are not hardened extremists, who operate in encrypted channels beyond the reach of any statute, but ordinary people. The present state of the UK, where expressing an opinion that gives rise to even mild offence may result in arrest, and does result in arrest, for tens of thousands of people per year, is a heavy price that this country has been paying for two decades.

The bill for the UK’s bad conduct now falls due. Overzealous enforcement of these laws has resulted in the arrests of parents complaining about their local school boards; Jewish citizens being threatened with arrest for looking too “visibly Jewish” in public; satirists of every stripe, from the absurd to the gender-critical, being arrested and charged for ideas that, only a decade ago, were commonly held; teenagers threatened with terrorism charges for reposting memes.

Then, in 2023, Parliament vested Ofcom, the UK’s Internet censor, with vast powers, and spent hundreds of millions of pounds building a purportedly globe-spanning enforcement apparatus, only to lose in the court of public opinion to – and be proven utterly toothless by – a cartoon hamster. The UK’s censorship regime does not work, does not deserve to survive, and this Model Bill – if enacted – would put it out of its misery.

“If the UK wanted to enact something like the First Amendment, what would the resulting statute look like?”

That is the question we seek to answer with the publication of the Model Bill. A requirement of such a law would, necessarily, mean that on the day of its enactment it would need to dismantle virtually all the UK’s censorship regime in one, swift strike.

It would have to, overnight, eliminate the British state’s power to police the opinions of the British people, because the total elimination of that power – all of it, all at once – is what is required in order to truly recreate the First Amendment, warts and all.

In publishing the Model Bill, and in proposing the extent of repeals we propose, the Model Bill’s authors ask readers one question, and one alone: do you want the UK to have a free speech right that is equivalent to the First Amendment?

That question has a binary answer: yes or no.

That is the Model Bill’s purpose. It does not seek to reach a consensus or an accommodation with pro-censorship forces in the UK’s political establishment. It does not call for incremental tweaks. In publishing the Model Bill, we seek to start a debate about whether the UK state, as a general constitutional principle, should retain the power to censor any Briton’s nonviolent expression on pain of fines and imprisonment, under any circumstances. We seek a debate about the UK’s censorship laws and practices, all of them, all at once. We seek that debate because we are confident that debate is one we will win.

The Model Bill is an attempt to create a First Amendment for the United Kingdom. It is organised in eight Parts and six Schedules. It establishes a broad statutory right to free expression, defines narrow exceptions, imposes duties on the State and essential service providers, creates new protections in employment and equality law, repeals most of the UK’s existing legal censorship architecture, and provides enforcement mechanisms for persons whose rights are violated by the State.

The Right to Freedom of Expression.

Part 2 establishes the core freedom. The right of any United Kingdom person to hold opinions and engage in lawful expression may not be violated by the State (s.4). That right extends to expression that is offensive, grossly offensive, insulting, abusive, shocking, blasphemous, indecent, or otherwise objectionable (s.4(2)). Section 5 states the principle plainly: there is no right in law not to be offended by the expression of others.

Section 6 defines the irreducible core of expression the Model Bill seeks to inviolably protect: where speech on matters of public interest, including politics, morality, philosophy, and religion, is made, no person may be convicted or sanctioned for such expression unless it falls within a defined category of unprotected expression and the conditions for liability are strictly met.

Unprotected expression.

Part 3 defines what expression is not protected.

The Model Bill adopts the American Brandenburg test for incitement: inflammatory expression is unlawful where it is directed to a specific audience, intended to produce imminent lawless action, and likely to produce such action (s.9). Advocacy, praise, or defence of unlawful conduct in general terms is not incitement. Beyond incitement, the Model Bill preserves criminal liability for perjury, contempt of court, unlawful threats, fraud, blackmail, offences under the National Security Act 2023 and the Official Secrets Act 1989, procuring or assisting crime, criminal conspiracy, defamation, and harassment (s.11), as well as any unlawful speech specifically designated by Parliament.

The Model Bill also amends the Protection from Harassment Act 1997 to clarify that a person does not harass another merely by posting content that the other person chooses to seek out, subscribe to, or monitor (s.10), avoiding situations such as those where citizens who choose to report on matters of public interest, such as the police, without actually contacting the subjects of the reporting, are then arrested under that Act.

Restrictions on State action.

Part 4 prohibits the State from censoring lawful expression directly or indirectly (s.13), including by imposing conditions on licences, funding, or benefits that require a person to adopt a particular political or ideological opinion. Section 14 bans non-crime speech monitoring, including the recording of lawful expression as a “non-crime hate incident,” and requires the destruction of any records that may exist in violation of that rule. Section 15 closes the outsourcing loophole: the State may not arrange for, fund, or procure censorship through third parties.

Compelled speech is prohibited: no public authority may require any person, as a condition of employment, education, professional licensing, or citizenship, to declare or affirm any political, moral, religious, or ideological belief, other than allegiance to the Crown (s. 16).

Intellectual discrimination.

Part 5 amends the Employment Rights Act 1996 to protect workers from detriment and employees from dismissal on the ground of lawful expression made in a personal capacity outside the workplace (Schedule 3). It also amends the Equality Act 2010 to make “lawful expression” a protected characteristic for the purposes of services, education, employment, and associations (Schedule 4), in keeping with the Equality Act’s protections for belief. If belief is protected, so too must be the expression of those beliefs outside the workplace.

The Model Bill also inserts a freedom from compelled expression into the Equality Act: nothing in that Act requires any person to express, endorse, or facilitate any message, opinion, or viewpoint, and a refusal to do so does not of itself constitute discrimination.

Essential services and the Internet.

Part 6 prohibits essential service providers – including banks, payment services, telecoms, and domain registries – from refusing or withdrawing services on the ground that a person has engaged in lawful expression (s.19). Again, this is something which would be unconstitutional in America but fits within the UK’s tradition of non-discrimination.

Section 20 will doubtlessly be one of the most controversial items of legislation in this proposal. The repeal schedule repeals the Online Safety Act in its entirety, and that entire regime is replaced with a provision modelled on Section 230 of the U.S. Communications Decency Act: no provider or user of an interactive computer service is to be treated as the publisher of content provided by another.

For such a short provision, the American experience shows that the decision to enact a Section 230-style rule is very complex. This decision was informed by a number of competing policy considerations, foremost among them the need for the UK to succeed economically in a digital and globalized world.

That does not mean that the Free Speech Act proposes to turn the UK into a digital wild west: unlike the effect of Section 230 on U.S. state criminal laws, UK criminal law is completely unaffected by this immunity, which applies to civil liability only. Like Section 230, platforms retain full discretion to moderate content they consider objectionable, and with a handful of exceptions, many major platforms already censor content that the UK finds illegal. Section 20 imposes one mandatory obligation on platforms: they are required to detect and remove CSAM, and render a report to law enforcement within 24 hours of obtaining actual knowledge of its existence (s.20(3)). Automated systems widely in use should make the takedown-and-reporting procedure nearly instant in most cases. Failure to adhere to this requirement is already punishable under existing UK criminal law.

So, the position on “online safety” in the FSA is reverted to that in January 2025. I lived in the UK as recently as 2017 and it wasn’t a post-apocalyptic anarchist hellscape; repealing the Online Safety Act in full, as this Model Bill proposes to do, won’t make it one. It will, however, end the UK’s exercise of censorship powers over political speech online, or requiring every person in the UK to dox themselves in order to use their iPhones. That seems to me like a good trade.

Enforcement and remedies.

Part 7 gives individuals a civil right of action against the State or essential service providers that act incompatibly with the Model Bill (s.21), with a presumption of costs in the claimant’s favour. Section 22 creates an anti-SLAPP mechanism: defendants in strategic lawsuits against public participation may apply for early dismissal, with mandatory costs and the possibility of damages if the claim is struck out. Section 24 annuls and vacates all convictions, cautions, and binding-over orders entered under repealed provisions where the conduct would constitute lawful expression under the Model Bill.

Repeals.

Schedule 1, Part 1 is the wrecking ball.

Schedule 1 repeals the Public Order Act 1986, the Public Order Act 2023, the Malicious Communications Act 1988, the Online Safety Act 2023, the Hate Crime and Public Order (Scotland) Act 2021, and the Obscene Publications Act 1959, in their entirety. It also repeals Section 127 of the Communications Act 2003, sections of the Terrorism Acts used to arrest non-violent protestors in the UK for political speech and Article 10(2) of the Human Rights Act 1998’s Schedule 1. It eliminates the strict liability rule for the press on court proceedings under the Contempt of Court Act 1981.

Schedule 2 re-creates the content-neutral public order offences of riot, violent disorder, affray, and unlawful assembly, only this time, these provisions are shorn of any content-based element; redefines breach of peace; and creates two new content-neutral communications offences that punish the conduct of threatening another person, regardless of the ideological orientation of the threat in question: (a) fear or provocation of violence and (b) threatening communications.

Schedules 5 and 6 are placeholders for applicable rulemaking in Scotland and Northern Ireland. We presume Labour will not be introducing this Model Bill so we have a few years to address the interaction between free speech and devolution.

Why this Model Bill? Why now?

When I wrote Sense and Sensitivity in 2020, my early prototype for a UK Free Speech Act was met with warm words, some curiosity, and no action.

The intervening years have made the case for law reform even stronger. The Online Safety Act was passed. Ofcom was set loose on the world. Scotland’s hate crime law came into force to widespread ridicule and even public opposition from the then-Prime Minister, Rishi Sunak (who, despite that opposition, did nothing in response). The summer of 2024 saw fresh speech prosecutions. The Online Safety Act’s entry into force in 2025 accelerated the decline.

My two co-authors, Michael Reiners and Elijah Granet, and I offer this Model Bill as a template, a menu of options, for any UK politician or citizen who wants to halt the collapse of Orwell’s country into Orwell’s fiction.

The Model Bill is not a radical proposal. The Free Speech Act is perfectly aligned with Britain’s liberal history and traditional British values, values that are sane and just, and values which the UK’s current speech rules do not align with. The radicals are the politicians, pressure groups, and companies who are either actively trying to turn, or quietly acquiescing to the transformation of, the UK into Censorship Island.

The establishment should have to answer for a legal regime that causes half the British population to self-censor out of fear and which arrests tens of thousands of people per year for simply speaking their minds. They cannot, because it is indefensible.

The Model Bill is not a sacred text. It is simply three lawyers’ vision for a free Britain. We expect readers of the proposal to suggest improvements, amendments, additions, deletions, or even to throw the Model Bill out and propose novel approaches of their own.

But, by way of reminder, you either want a UK First Amendment, or you don’t. If you think that our Model Bill confronting these issues head on, as we have done, is too controversial, we invite you to explain to the British public why the status quo is not.

Whether the UK chooses to go where the Free Speech Act leads, or whether it follows another map entirely, is a matter for the British people to decide.

Preston J. Byrne is a co-author of the Free Speech Act 2026 Model Bill. He is a dual-qualified English solicitor and U.S. attorney-at-law, Senior Fellow of the Adam Smith Institute, and Managing Partner of Byrne & Storm, P.C.

He is counsel to 4chan in its federal lawsuit against Ofcom in the Federal District Court of the District of Columbia, and to every U.S.-based enforcement target of the UK Online Safety Act in 2025. He also represents American enforcement targets of Internet censors in the European Union, Australia, and Brazil. He is also the creator of the GRANITE Act foreign censorship shield law concept for the United States, and co-author of HB 70, the Wyoming GRANITE Act, which passed the Wyoming House of Representatives 46-12 in February 2026 and is currently under consideration by legislators in several states.

The UK Free Speech Act 2026 Model Bill is published alongside this essay and is available for download at adamsmith.org.

Early feedback from lawyers on the Adam Smith Institute’s UK Free Speech Bill

In advance of Wednesday’s expected publication of the UK Free Speech Bill proposal, I wanted to offer some clarifications ahead of time to those of you who have read the proposal.

For those of you who haven’t, this blog post will make very little sense until the full bill is released.

A note on my co-authors and other contributors

Although I came up with the concept for the Free Speech Act and have been the frontman for the band, the Model Bill has two other co-authors: Michael Reiners and Elijah Granet.

Michael Reiners is a Cambridge-educated English lawyer who writes prolifically on UK constitutional and political issues. His role in preparing the draft was to beef up my initial outline proposal and provide a lot of the operative language. He was also responsible for introducing the Equality Act and Employment Rights Act concepts in the document.

Elijah Granet has dual English and American legal training with professional experience consulting in legislative drafting and analysis. He provided valuable input on First Amendment principles and UK constitutional principles, as well as Parliamentary-style drafting and conducting an OPC-style proofreading. He was also solely responsible for the beautiful formatting of the Model Bill as-published.

Their input was invaluable, is woven everywhere into the bill (which was drafted collaboratively using Google docs) and without it the Model Bill would never have been completed and certainly wouldn’t be as good as it is.

I also thank the several members of the U.S. and UK free speech bars and free speech activist communities for their input on the Model Bill during the fairly compressed drafting process over the last several months, which was gratefully received and which resulted in a number of critical changes during the drafting process.

Section 4 (generally)

I received some specific comments on Section 4, but wanted to add a general comment of my own about Section 4 before proceeding to the rest of the comments that came in.

Section 4 is really the only language that I would regard as “sacred text” in this bill.

I have not written a policy paper or a detailed analysis or a law review paper about this. The only way I can explain why this is worded the way it is is that I have been active in the UK free speech field and working on this definition for fifteen years – and working on it in public since 2020 when I published a prototype version of Section 4 for the first time, then as Section 1(1) and 1(2) of the “UK Free Speech Act 2021.”

When I see a case of UK political censorship that should never have happened, and I try to resolve it in my own mind, the overwhelming majority of cases resolve correctly with this formula. “Resolve correctly” here means “a free speech lawyer working in England would probably agree with the outcome.” For this reason this language is, in my opinion, the one that makes the most sense to replace Article 10 for United Kingdom persons.

Putting it another way: Section 4 is the load-bearing structure which supports the rest of the Model Bill.

Definition of “State” and “public authority”

We do not want Human Rights Act 1998 caselaw to carry over to this Bill, so we chose the term “State” for most of the operative provisions.

“Public authority” is a separate term which does not include the Army or the CofE, and is extensible via Schedule 7.

It is not perfect, but we had a choice to make: get this bill out there in time for free speech reform to be an issue in local elections, or quibble over definitions. We chose the former. This is part of the outer boundary limit of any future UK free speech right that will need to be negotiated and decided among later drafters. Graham Smith has already helpfully picked apart this definition and we’ll look forward to further comments along these lines from him and others, assuming the concept gets picked up and there’s a drafting process to move forward with somewhere that has a reasonable chance of resulting in enactment in the next decade.

Definition of “United Kingdom person.”

I don’t actually personally agree with the definition we used here, even though I wrote a lot of it. If it were up to me, I would have cloned the First Amendment – anyone within the jurisdiction of the UK benefits from the right, including, potentially, foreign citizens on U.S. servers posting anonymously (see, e.g., In Re DMCA Section 512(h) Subpoena to Reddit, Inc. or In re Subpoena to Reddit, Inc.)

We did not explicitly carve out anonymity in our proposal; I think it is likely implied by the definition of “expression” and the right not to be engaged by compelled expression, but our failure to address it up front was an omission. This is an issue iterations of this bill would need to address or, if not addressed, that the judiciary would likely decide. How to protect anonymity whilst proving entitlement to the right is an open question under the bill’s current language.

The reason for this language selection is that, while I am a U.S. libertarian, the current major opposition parties in the UK who are likely to form the next government – Reform and the Conservatives – are not libertarian parties. As the Free Speech Act is a thought experiment seeking to establish what a British First Amendment might look like, this definition was thus tuned to a definition that the probable successors of the Labour Party in government would be most likely to accept.

This definition leaves out one category of claimant: it protects (a) UK citizens, regardless of their location; (b) individuals who are UK resident, regardless of their location; and (c) individuals who are lawfully present in the United Kingdom, by whatever means (citizenship, work visa, tourist visa, asylum/refugee status).

If you are not lawfully present in the United Kingdom on any basis, and are not a citizen or lawfully ordinarily resident, you do not get the turbocharged rights under the Model Bill. This is a significant departure from both the First Amendment and Strasbourg. Given that Article 10(1) is not repealed by this Act, however, the Article 10(1) right remains applicable to individuals who are not lawfully present in the UK. That right is enhanced by the Article 10(2) repeal. So the position of someone who is unlawfully present in the UK under the UK Free Speech Act regime would still be better than it is currently under Strasbourg jurisprudence, which everyone else would still get, but would not benefit from full UK Free Speech Act rights, which requires lawful presence.

The UK Free Speech Act, if enacted, would be one of the most powerful and assertive free speech regimes in the Western world. In some respects it protects expression in domains the First Amendment prohibits, particularly in the employment context. I am not married to this definition but I did not think that UK political parties would be keen on extending the protections of such an expansive regime to the entire world and would want to limit it to Britain’s political community, which remains nonetheless very broadly defined.

This principle carries over to corporate speech. Note that foreign corporations and organizations do not benefit from Free Speech Act protections, although their UK establishments do – see 4(d) through (h).

This Model Bill is, of course, just a proposal. What to propose or enact is a matter for politicians to, in their infinite wisdom, decide. In the meantime we offer some (very rough) amendments that might patch the anonymity issue and the “United Kingdom person” issue to the satisfaction of a more libertarian-minded legislator below.

Seeing as we’ve published the Free Speech Act to the world I should imagine a lot of other people will have a lot of other opinions on the subject. If they like the Model Bill I trust they will pick up where we left off.

Section 4(3): the irreducible core of freedom of expression

I got a lot of comments on this one, from both U.S. and English free speech lawyers. Nobody really understood it.

The language is intentional. The intention of Section 4(3) was to create and define an irreducible core of freedom of expression, the inner limit of the right, subject to specific and express exceptions, the outer limit of which will, in all probability, be defined by the judiciary in something more than its usual judicial review function but something less than the U.S. judicial review function.

American practitioners, in particular, objected to the choice to make certain content-based categories of speech privileged over others. The reason for this is simple: the judiciary needs time and cases to develop a doctrine around the outer limits, but any solution to the UK’s free speech emergency needs to address the inner limit on the day of enactment, and define that limit so that only an Act of Parliament targeting that inner limit, rather than an overly censorial first-instance judge or public authority, can override it.

The First Amendment has had 250 years to evolve. The Model Bill, or one like it, if enacted, will have to function on the day of enactment. That’s a lot of catching up to do. One bill can do a lot of catching up, and can wipe away a lot of the Strasbourg-style balancing tests, but one bill cannot do it all and it would be folly to try. (I have seen other, admirable attempts to do this and it was decided we would not follow that approach.)

If a bill like the Model Bill were enacted, the UK’s free speech doctrines would take many decades to evolve and mature after that. Thus, we enumerate specific categories of speech for special protection, both in Section 4 and Section 6, but qualify those lists (which would otherwise be restricted by the inclusio unius rule of construction) with “without limitation” language in order to communicate to the judiciary that this is the inner limit and not the outer limit.

Section 8(2): Should the question of whether speech falls within an unprotected category be treated as a question of law decided by judges, or a question of fact decided on the balance of probabilities by magistrates or juries?

Good question! I don’t know the answer. That is something that, if the Model Bill is adopted by any of the major parties, those parties will have to decide for themselves.

Section 9 – Direct Incitement

“How would this Model Bill deal with the Lucy Connolly situation?” The answer is contained in Section 9, which imports the U.S. Brandenburg test, with minor modifications. Connolly’s speech would be run through the three-part Brandenburg test (the original has two steps rather than three) and a judicial fact-finder would reach a determination.

We did not make this a standalone offence, but rather a qualifier of the free speech right – the test would presumably be used in conjunction with the reformulated public order and communications offences, or as part of the elements that need be proved in a prosecution of the inchoate offences of encouraging or assisting the commission of crime in Sections 44-46 of the Serious Crime Act 2007. It may be better to insert this into the Serious Crime Act 2007 as an amendment. I leave that decision to future drafters.

Sections 16 and 17: Equality Act and Employment Rights Act.

This is more language I disagree with that found its way into the final Model Bill.

The first proposal I can find to do something like this was written by prominent election law barrister Francis Hoar in his 2021 paper for the Reclaim Party, In Protection of Freedom of Speech. The language found its way into this Model Bill via my co-author Michael Reiners.

I oppose this personally, as did several U.S. First Amendment lawyers I shared it with, because it allows the government to intrude into employment relationships on the basis of protected speech – which itself interferes with the employer’s rights of free association. If it were up to me, I would abolish “belief” as a protected characteristic in the Equality Act, as I have written about before in my short essay UK Free Speech as a Sword, Not a Shield.

But creating a document that reflects my personal preferences is not the point of this exercise. The point of this exercise is to liberate the United Kingdom, and in order to do that the co-authors had to ask ourselves what a British First Amendment might look like.

As Toby Young wrote in 2024, “a British First Amendment wouldn’t save free speech” because a clone of the First Amendment is not possible under the UK’s constitutional structure. His short 2024 essay on this topic directly influenced the drafting of this Model Bill; the UK is not America and it never will be. We also had to work within existing political constraints, one in which Reform and the Conservatives, not U.S. libertarians, are our target audience. So the language stayed in.

There may be good, uniquely British reasons to depart from First Amendment doctrine in this area. In some parts of the UK, parts of Wales and Scotland, the government accounts for nearly 2/3rds of GDP, meaning private employers may be more susceptible to government pressure and the additional protections here are needed.

Section 20: Novel platform immunity provision

This is directly inspired by Section 230 and that choice is intentional.

The drafters think, on balance, that the U.S. equivalent, Section 230 of the Communications Decency Act, was the correct policy choice to make. The absence of a federal system (more or less) in the UK means that there is no need to disapply state criminal law, hence the immunity is expressed to be civil only. There is plenty of material from the United States to inform UK policymakers as they consider Section 20.

Our argument for approaching the “child safety” issue in this way is that Ofcom, as an Internet censor, is not qualified to do the real work required to promote “safety” – since that work is kinetic and involves putting predators and other dangerous people in handcuffs and, in due course, in prison. The mandatory takedown paired with mandatory and prompt law enforcement agency reporting mechanism accomplishes the “safety” objective far more effectively than any censorship law, including the Online Safety Act, does.

Section 21: The civil proceedings mechanism may not produce substantial damages awards.

Also a fair point. We didn’t want to go overboard on damages as the primary relief a claimant can obtain in these cases – the desired relief – is injunctive or declaratory relief. This Bill is not meant to be a boon for the plaintiffs’ bar; free speech defense is not a very lucrative business. Ask me how I know.

This question is something that, if the Model Bill is adopted by any of the major parties, those parties will have to decide for themselves.

Section 26, Sch 1 Para 1 repeal of Article 10(2):

We fully expect the human rights bar to go nuclear over our choice to repeal Article 10(2) of Schedule 1 to the Human Rights Act 1998 and subordinate Strasbourg jurisprudence to it.

Britons have free speech rights. This Model Bill protects those rights. Strasbourg-style interest-balancing exercises do not. Article 10 of the European Convention is a censors’ charter and it is time for it to be abolished for British citizens. This Bill does so, intentionally.

“But international law!”

Parliament is sovereign, it does what it will. First year law school stuff. If Parliament makes that call, there is nothing a bunch of ECtHR judges in Strasbourg can do about it.

Schedules 1 and 2

The table of repeals in Schedule 1 was selected on the basis of one criterion, and one alone:

Whether the law, as-written, is compatible with the freedom created by Section 4.

I suggested most of these repeals. Elijah raised the Obscene Publications Act, inclusion of which I agree with. Michael suggested the Senedd Cymru (Member Accountability and Elections) Act, with which I also agree. Section 4 creates a robust free speech right. If you accept that the right should follow the contours of Section 4, it follows that any law that violates that right should be repealed.

We looked at laws in common use, with a particular focus on the criminal law, and concluded that the laws in our table violated Section 4’s free speech right.

We then re-enact a number of those laws in a manner that complies with Section 4 in Schedule 2. This includes,

  • in Part 1 of Schedule 2, re-enacting riot, affray, violent disorder and unlawful assembly;
  • in Part 2 of Schedule 2, re-enacting a replacement for the “stirring up racial hatred” and threatening offences as “fear or provocation of violence” (para. 5) and a new communications offence derived from Section 181 of the Online Safety Act 2023, “threatening communications” (para. 6); and
  • in Part 3 of Schedule 2, providing a statutory definition for “breach of peace” that codifies the rule from Beatty v. Gilbanks, [1892] 9 QBD 308, and operates to abolish the heckler’s veto in a public order context.

The threatening communications offence, in particular, that we propose is significantly broader in one crucial respect than most of the individual offences under the Public Order Act because it covers threats made “by any medium of expression,” including speech and conduct. Where threatening behavior is concerned, therefore, this compresses part of Section 4, and all of Sections 4A, 5, 17, 18, 19, 20, 21, and 22 of the Public Order Act 1986, and Section 181 of the Online Safety Act, Section 127 of the Communications Act 2003 and Section 1 of the Malicious Communications Act, into a single offence which, shorn of any content-based component and dealing only with the conduct or nature of the speech, (a) complies with the First Amendment of the United States Constitution and (b) more importantly, also complies with the proposed Section 4 of the Free Speech Act.

The replacement is simpler, broader, and more content-neutral than the regime England and Wales currently has.

Yes, our proposal means that being a bigot or an asshole in public or on the Internet will not result in a prison sentence. Our proposed regime is not a free-for-all. Under our regime, if you threaten or harass someone, for any reason, regardless of the content of the ideas expressed, you still go to prison. Under our approach, the state doesn’t police opinions, it polices conduct. As it should.

Another plus point from a wholesale repeal of the Public Order Act 1986 is that the caselaw is so polluted with Strasbourg-style balancing tests that take what would, in America, have been constitutionally protected expression and nonetheless result in convictions upheld on appeal (Norwood, Hammond, Abdul, etc.) that it might be better, if we’re resetting, to not amend the Public Order Act 1986 at all, jettison the old caselaw that will be abolished under Section 4, and just burn the thing down.

The Hammond case, in particular, has loomed large in my mind for well over a decade: it is the polar opposite of pre-HRA Redmond-Bate on identical facts. I read those cases and that tells me there’s absolutely nothing about the Public Order Act 1986 that deserves to survive, unless you’re prepared to argue how you can have free speech in a world where Hammond was correctly decided and Redmond-Bate was not.

If we missed anything that folks think should be ported over from the Public Order Acts or the Online Safety Act that is up to future drafters to pull across on their own versions of something like the Model Bill.

To be consistent with the approach taken by the Model Bill, such offences should be reformulated so that there is no possible interpretation of the offence which would conflict with the right in Section 4. Part 3 of the Public Order Act might be difficult to repeal from an optics standpoint, but if you want a First Amendment-style right in the UK these laws must be repealed and reformulated.

Possible approaches include, e.g., adding hate crime enhancements to the content-neutral public order offences where bias-motivated crime attracts heftier penalties. What you cannot do while also having a First Amendment-style rule, is punish speech more heavily because of the content of the speech, because if a bigot doesn’t have the right to hold and express an opinion, nobody does.

The question of whether a nation’s laws truly permit free speech is tested at the extremes, not at the center. Free speech is judged by whether it protects the worst opinions, not the best ones. That is not the price of liberty; that is liberty and if the people who hold the worst opinions get it, the rest of us get it too.

If you’re not prepared to accept that, you’re not actually in favor of free speech. You’re in favor of the government punishing people for expressing opinions. That is not free speech, it never has been, and it never will be.

Schedule 1: Repeal of certain provisions of the Terrorism Acts.

This is controversial. I knew it would be controversial when I wrote my first “UK Free Speech Act” proposal in 2021, so I chickened out and didn’t mention it.

I first suggested repealing it publicly in September. I knew it would be controversial then, but nobody noticed that Substack post so nothing happened. Nothing between then and now has changed my assessment on that point.

It made it into the final Model Bill because the time for chickening out on free speech in the UK is past. The sight of grannies and vicars on Parliament Square getting arrested for nothing more than peacefully holding signs in a protest zone, simply because of the ideas the signs expressed, means that in order for this exercise to not be intellectually dishonest, repeals to the Terrorism Acts had to find their way into the Model Bill.

There changes are, unfortunately, necessary ones if the UK is to restore its free speech rights. They are necessary because Parliament granted the police enormous powers and the police are now abusing those powers to suppress speech.

In the United States, these laws would be unconstitutional, facially and as-applied. Full stop. Speech crimes like the encouragement offences, expressing support for proscribed organizations, or possessing “articles” useful in terrorism have led to academics like the Nottingham Two, or Palestine Action supporters protesting peacefully on Parliament Square, being arrested and threatened with decades in prison.

The process of proscription itself would be, in America, both unconstitutional as punishing expressive association and an unconstitutional bill of attainder, which America abolished in 1788.

If you want to regulate encouragement of terrorism in a free speech-protective manner, the way to do this is to look to the American approach of terrorism enhancements, terroristic threats, the law of conspiracy or “material support,” which will all gybe with Section 4, draft them in a content-neutral fashion, and chuck them into Schedule 2. If you want to deal with terrorist associations, drop the proscription and focus on beefing up terrorist conspiracy statutes along the lines of 18 U.S.C. §§ 2339A, 2339B, and 2332b.

The laws we targeted for removal from these Acts are used to target nonviolent political speech. For the UK to restore free speech, these laws have to go or be heavily amended. If you want to replace them, go for it, but keep in mind the right in Section 4.

It’s a hard conversation to have, but the time to have it is now. Incrementalist hedging has not been a winning strategy for the UK free speech movement in the past and it will not be one in the future.

Schedule 1 Part 3: Outraging Public Decency

Elijah suggested this repeal and I agree with it. Some practitioners questioned this repeal. We feel that truly criminal conduct in this area is adequately covered by the Sexual Offences Act, although amendments to the SOA may be warranted to broaden it.

There is no question in our mind that the common law offence violates free speech rights. The recent arrest of adult performer Tia Billinger for performing an “obscene” gesture with her hand in protest of her arrest and deportation from Bali, for example, that would have been unambiguously constitutionally protected in the United States makes that clear enough.

As mentioned, any law we regarded as inconsistent with Section 4 was repealed. No exceptions. This Bill was a volunteer effort by three lawyers in our free time; given time and resources, the recommended repeal or abolition list would have become longer, not shorter.

If someone wants to draft a replacement for any of the laws we have proposed to repeal, they are welcome to, and we would not have any objections to such a replacement provided that the replacement consists with Section 4.

Schedules 5, 6: Reserved matters for devolved administrations

Devolved administration issues are going to be an issue. We do not offer a solution here. For a future government to resolve, but Parliament has the power to unilaterally grant free speech rights to the entire UK and, in my reasoned opinion, it should use it.

In my opinion, free speech is too important an issue to be held hostage to the politics of devolution. That is, of course, a matter for politicians to decide, rather than me. This Model Bill is just my thinking on the topic and anything in it, barring Section 4 (and probably also barring, to a lesser extent, any major reworking of Sections 6, 13, 15, and 16) can be tweaked and swapped out without completely breaking the Model Bill.

In any event, if there’s the political will to enact a Free Speech Bill, there’ll probably also be the political will for broader constitutional reorganization. I don’t think too many people in Scotland would be unhappy to see the Hate Crime and Public Order (Scotland) Act 2021 go the way of the dinosaurs.

Communications Act 2003

We deliberately sidestepped the question of Ofcom’s authority to regulate broadcast communications and the BBC, although there is a free speech dimension in how Ofcom performs that function.

We considered using the repeal table in Schedule 1 to perform some hypothetical emergency surgery on Ofcom’s content regulation powers, but ultimately decided against it in this version of the proposal because this is a “culture war”-coded issue in the UK and wouldn’t really address the core free speech emergency – stopping the practice of police power from being used to control discourse.

For now, it might be better for a government-in-waiting that wants to proceed with something like the Free Speech Act to make this the topic of a separate bill which is designed to interface with the Free Speech Act.

HRA 1998, Schedule 1, Article 8 and the Defamation Act 2013

For broadly similar reasons, we sidestepped the interface of Article 8 and UK defamation law with this new proposed free speech right, although there are definitely touchpoints and a political party seeking to bring forward something like the Model Bill will need to decide where to draw the line.

The primary purpose of this Model Bill is to terminate the practice of criminal arrests for what ought to be protected expression. Defamation and privacy is a separate topic that is best left to a political party’s policy team to decide where to draw the line with another Act of Parliament that seeks to amend that other part of the Human Rights Act.

Alternative architecture: some practitioners suggest that the bill’s objectives could be achieved more efficiently through a shorter statute.

We look forward to seeing those proposals!

The purpose of our Model Bill was to start exactly these kinds of conversations.

On the So to Speak podcast with Nico Perrino

I was on FIRE’s podcast, “So to Speak,” with Nico Perrino talking about foreign censorship. Recorded last week, out today.

“I never thought the lion would eat MY face!”

Ofcom seems to be pivoting away at light speed from smaller targets and going after the U.S. majors, like Google.

I had wondered why I got 65 hits on my site last night from Council Bluffs, and today we know: Ofcom and the UK Information Commissioner’s Office opened new files on Facebook, Instagram, Snapchat, TikTok, YouTube, Roblox and X.

One company wasted no time trying to throw my clients, Ofcom’s current targets, under the bus:

This is not a winning strategy anymore.

A message to my colleagues in Big Tech:

That wall is actually pretty easy to use: Ofcom can threaten an American company all it wants, but the record shows that threats cannot translate into enforcement across America’s borders. Using the wall is as simple as saying “no.”

We are treating Ofcom’s officious censorship demands accordingly:

The First Amendment is sitting there on the ground, waiting for American technology companies to pick it up and wield it. Will they have the courage to do so?