16.1 What freedom of speech actually is

Legal protection, not entitlement

Freedom of speech protects you against certain kinds of interference. It is not a promise that anyone must hand you an audience or a platform.

In the UK, the protection most often cited is Article 10 of the European Convention on Human Rights, brought into UK law by the Human Rights Act. It protects the right to hold opinions and to receive and impart information and ideas. It also allows restrictions that are "necessary in a democratic society", such as laws against incitement to violence, harassment, or disclosing certain confidential information.

How that plays out day to day:

• A local council cancels a peaceful public meeting because it dislikes the organiser's views: that can be challenged legally.
• A newspaper refuses to print your letter to the editor: no legal entitlement to force it.

The right limits what the state can do. It does not compel private actors to host or amplify you.

Myth: free speech means freedom from consequences. Reality: it does not.

• Publish inaccurate claims about a neighbour and you may face a defamation claim.
• Leak an employer's confidential client list and you may be sued for breach of contract.

These are not "exceptions" that break the principle. They are the normal boundaries where the law balances speech against other rights such as privacy, reputation and safety.

Negative rights and positive rights

Quick definitions:

• A negative right requires others to refrain from interfering.
• A positive right requires others to take action.

Freedom of speech is primarily a negative right: the state should not censor or punish lawful expression. By default it does not oblige the state to give you a printing press, a broadcast licence or a guaranteed place in a public debate.

Some positive obligations do exist, but they are narrower and often policy-driven rather than absolute. For example:

• public bodies might have to act fairly and consider a range of views in certain consultations;
• a publicly funded broadcaster may have duties around impartiality and access.

These create opportunities to speak, but they are not an open-ended entitlement to be heard everywhere.

This helps make sense of access disputes. A university hosting a public lecture may have obligations around free expression on campus, but can still set rules on safety, security and scheduling. A private conference rejecting a speaker is usually a contractual and reputational matter, not a constitutional one.

Speech versus reach

Being able to speak and being able to reach people are different things. Speech is the act of expressing or publishing. Reach is distribution: how far a message travels, who sees it, and how it is ranked or surfaced.

In a digitally monitored world, reach is shaped by technology as much as by law. Two examples:

• A community group publishes a statement on its own website. That is speech. Whether it shows up in search, gets shared on social platforms or is recommended by an algorithm is reach.
• A journalist posts a story on a personal blog without interference. Its reach may still be limited by platform policies, advertising rules or recommendation design.

Is limited reach censorship? Sometimes. Sometimes it is a by-product of unrelated design choices, such as filters that prioritise trusted sources during a public health emergency, or ranking systems that downrank spam and misleading content.

The risk is that opaque or badly designed systems can suppress legitimate speech without anyone intending it. That is not purely a technical problem; it is a governance problem.

The fixes are mostly practical rather than legal:

• Use several channels rather than relying on one platform.
• Keep copies of your work and publish on sites you control.
• Learn how moderation and ranking work, so you avoid accidental triggers such as repetitive posting that looks like spam.

None of this guarantees reach, but it lowers the chance of your speech vanishing because of avoidable technical or policy quirks.

No First Amendment: why the UK is more restrictive than people assume

A lot of free-speech talk is borrowed, often unconsciously, from the United States. There, the First Amendment gives unusually strong protection against government restriction of speech. It is not absolute, but its structure and case law differ sharply from the UK's.

Article 10, given domestic effect by the Human Rights Act 1998, is a qualified right. Any interference must:

• have a legal basis;
• pursue a permitted aim;
• be necessary and proportionate in a democratic society.

Courts must balance expression against interests such as public safety, preventing disorder or crime, privacy and reputation.

So some expression protected in the US may be unlawful in the UK. But it is equally wrong to say offensive speech has no protection here. Article 10 protects expression that offends, shocks or disturbs; context, intent, statutory elements, necessity and proportionality all matter.

Parliament can create communications and public-order offences, but police, prosecutors, public bodies and courts remain subject to legal safeguards. The useful lesson: the boundary is neither absent nor wholly subjective. It is fact-specific, and cannot be boiled down to an American slogan or a list of forbidden opinions.

The specific UK laws that limit speech

"Permissible restrictions" gets much clearer once you know the actual statutes. Speech is governed by a patchwork, not a single UK code:

• Section 127 of the Communications Act 2003 still covers messages sent over a public electronic communications network that are grossly offensive, indecent, obscene or menacing.
• Section 1 of the Malicious Communications Act 1988 (England and Wales) still covers certain indecent or grossly offensive communications sent with a purpose of causing distress or anxiety.
• The Online Safety Act 2023 (England and Wales and Northern Ireland) replaced older false-message and threat limbs from 31 January 2024 with more specific false and threatening communications offences. The false-communications offence requires knowledge of falsity, an intention to cause non-trivial psychological or physical harm to a likely audience, and no reasonable excuse.
• Those Online Safety Act offences do not extend to Scotland, and Northern Ireland also has its own Malicious Communications Order.
• Harassment, stalking, contempt, defamation, confidentiality, terrorism and public-order law also differ by jurisdiction and facts.

These laws are real, but arrest, charge, prosecution and conviction are different stages. Context and mental state matter, and prosecutors must consider whether interfering with Article 10 is necessary and proportionate.

Who decides depends on where you are. The Crown Prosecution Service guidance describes the England-and-Wales approach; Northern Ireland has separate Public Prosecution Service guidance; a Scottish case must be assessed under Scots law. At protests, further rules now include the Crime and Policing Act 2026 provisions explained in 16.13.

The practical point: assess the exact communication, place and setting. Do not assume that causing offence alone is criminal, or that calling something a joke makes it lawful.

Primary sources: the Online Safety Act 2023 explanatory notes on territorial extent, Crown Prosecution Service communications-offences guidance, updated 29 June 2026, and the Public Prosecution Service for Northern Ireland's electronic-communications guidance.

None of this is a reason to fall silent, and it is emphatically not legal advice for any particular situation. It is why the rest of this chapter exists:

• lawful and safe are not the same thing (developed in 16.10);
• the boundaries are vaguer than people expect;
• the sensible response is informed care about audience, channel and identity, not reckless candour or frightened silence.

Knowing which laws exist is the first step to making those judgements deliberately rather than by accident.

Courts versus platforms

Courts decide whether speech is lawful. Platforms decide whether it is allowed on their services. Related, but not the same.

• A court may find a statement legal, while a platform still removes it for breaching community rules.
• A platform may allow content that is lawful but controversial, while users and advertisers push it to change course.

Real-world examples:

• A charity shares graphic images to raise awareness of abuse. A court would likely protect that as lawful expression; a social media platform may restrict it under its rules on violent or disturbing content.
• A local activist is cleared after a police investigation, yet a platform keeps their post removed because it was mass-reported and flagged as "misleading" during a heated news cycle.

The risk is twofold:

• Platform decisions can be inconsistent or poorly explained. Appeals often exist but may be slow or opaque, especially for smaller accounts.
• Legal protections are hard to enforce against private companies, particularly when decisions are automatic and at scale. You can complain, use formal reporting tools and in some cases pursue legal remedies, but it is rarely quick.

So build for resilience:

• Keep an off-platform archive of your work.
• If you rely on a platform for income or community, diversify and build direct contacts such as mailing lists or RSS feeds.
• When publishing sensitive material, document sources and context so you can contest removals with evidence rather than argument.

None of this removes the power imbalance, but it reduces the cost of a bad decision.

In the UK, cultural expectations around civility and harm often show up in platform policies, even where the law itself is less restrictive. A lawful comment may still be removed if it looks like harassment or encouraging hostility, especially in local community spaces.

Knowing those norms is not self-censorship. It is knowing where friction is likely and choosing the trade-offs that make sense for you.