GP / Blog / Politics & Governance

Issue forecast

Free speech in the online age: where the line sits between expression and harm

Free speech was last set in 2020, before the platforms became the main public square. The live question now is not whether speech matters but whether speech at the speed and scale of the internet can be as free as the principle once assumed.

Theme · Politics & GovernanceLast set at A-Level · 2020 (on whether a free society should have no restrictions on freedom of speech)Foresight, not a prediction

In short

A free-speech question rewards a student who can defend the principle and still think about its limits. 'All speech should be free' and 'speech that harms should be curbed' are both too blunt to win. Argue where the line sits, why it sits there, and use real platform regulation rather than abstract worry about censorship.

Why this could come up now

Free speech and its limits are a recurring strand. It appeared in 2017 on whether regulating the press is desirable and in 2020 on whether a free society should restrict speech at all. The principle is permanent; what changes is the technology that tests it.

The pressure now is concrete and dateable. The EU began enforcing its Digital Services Act in earnest in 2025 and 2026, and Singapore has built a stack of laws, POFMA and a new Online Safety Act among them, that govern online falsehood and harm directly. An examiner reading the news sees the free-speech debate moving from theory to live regulation.

Framed honestly: this is foresight, not a forecast of wording. A speech-and-regulation stem recurs, and its online form is the freshest version, because it forces the old principle to meet the new medium.

201420172019202020232025

set at A-Level most recent appearance. Recurs every few years through the free-expression strand; last set in 2020, before the platform-regulation wave.

What an essay on this would test

These questions test whether you can defend a principle without absolutising it. Almost everyone agrees free speech matters. The arguable claim is whether it should be near-total, or whether some limits, on incitement, defamation, deliberate falsehood, protect the very debate free speech is meant to serve.

They also test whether you can localise. Singapore's approach to online falsehood is firmer than the liberal Western default, so a Singaporean student who argues purely from an American free-speech absolutism will read as borrowed. The examiner wants you to reason about your own context.

Operative angles

no restrictions: a near-absolutist position the question often invites you to test, not simply endorse
harm versus offence: a crucial line, since speech that harms is a stronger case for limits than speech that merely offends
online: scale, speed and anonymity change what speech does, which is why the medium is part of the argument

How to answer it: two ways in

Two distinct, defensible routes through the question. A strong script commits to one and uses the other as the concession it answers, rather than sitting on the fence.

The principle: near-absolute

Speech curbed is debate diminished

Free speech is the precondition for every other freedom, and once a state can decide which speech is too harmful or too false, it holds a power that history shows is rarely used only against the genuinely dangerous.

The cure for bad speech is more speech: open argument exposes falsehood better than a censor does.
Whoever holds the power to ban 'harmful' speech decides what counts as harmful, and that line tends to creep.
Unpopular and offensive speech is exactly the speech that needs protection; comfortable speech never did.

Worked exampleWhen the EU fined X 120 million euros under the Digital Services Act in late 2025, the company challenged it in court as an attack on speech, a reminder that even well-intentioned regulation of online expression is contested as overreach by those it constrains (European Commission / reported court challenge, as of 2026-06).

The limit: harm at scale

Some limits protect the debate itself

Speech online is not a town-square pamphlet but a weapon that scales instantly, so a free society can defend reasonable limits, on incitement, targeted falsehood and coordinated harm, precisely to keep public debate usable.

Incitement, defamation and deliberate falsehood already sit outside free-speech protection in most democracies, so the question is where the line moves, not whether one exists.
Algorithmic amplification means a lie now travels faster and further than any correction, changing what 'let it be answered' can achieve.
States that judged online harm serious have legislated rather than waited, which suggests the costs are real, not imagined.

Worked exampleSingapore governs online falsehood directly through POFMA and added an Online Safety (Relief and Accountability) Act in late 2025, treating some online speech as a harm the state should be able to act against rather than leaving it to the marketplace of ideas (Singapore government, as of 2026-06).

The fuel: stats, facts and examples

120m euros

the EU's first major Digital Services Act fine, on X in late 2025, now under legal challenge

Source: European Commission, reported by IAPP and others · as of 2026-06

up to 6%

of global turnover, the maximum DSA fine for a platform that fails to comply

Source: EU Digital Services Act · as of 2026-06

Nov 2025

when Singapore's Parliament passed the Online Safety (Relief and Accountability) Act, adding to POFMA

Source: Singapore Parliament · as of 2026-06

Facts worth deploying

The EU has declared 2026 the year of Digital Services Act enforcement and now staffs a dedicated team to police platform compliance, so online-speech regulation has moved from drafting to active use.Source: European Commission, as of 2026-06

Singapore's POFMA lets ministers order corrections or removals of online falsehoods, and 2025 saw it used and a website blocked for non-compliance, giving a Singaporean essay a real, local regulatory model to weigh.Source: POFMA Office / news coverage, as of 2026-06

Most democracies already place speech such as incitement, defamation and child sexual abuse material outside free-speech protection, so 'no restrictions' is closer to a thought experiment than to how any free society actually works.Source: general legal consensus, as of 2026-06

FAQ

Is it safe to defend free speech strongly in a Singapore exam?

Yes. GP rewards a well-argued position, not a politically safe one, as long as you engage the counterarguments honestly. You can defend robust free speech and still acknowledge why Singapore regulates online falsehood. The essay is about the quality of the reasoning, not the conclusion.

Should I use American or Singaporean examples?

Use both, and note that they sit at different points on the line. American near-absolutism and Singapore's firmer regime are a built-in contrast that lets you argue where the line should sit rather than assuming one answer. The contrast is the argument.

How do I keep this from becoming a misinformation essay?

Misinformation is one case, not the whole question. Free speech also covers offence, dissent, art and protest. Use falsehood as your sharpest example of harm, then widen back to the principle, or the essay narrows to a single sub-topic and loses the bigger argument.