Freedom of speech was once imagined as a battle between the citizen and the state.
The old picture was clear. A government censored a newspaper. A ruler banned a book. A police force arrested a dissident. A court punished a critic. The threat to speech came from public power, and the defence of speech required limiting that power.
That picture is no longer enough.
Today, speech is pressured from two directions at once.
The state still matters. Governments censor, prosecute, intimidate, block websites, shut down the internet, misuse national security laws, pressure journalists, demand takedowns and punish dissent. But platforms now matter too. Social media companies decide what is amplified, demoted, labelled, removed, monetized, demonetized, recommended, hidden or made viral. Algorithms can give a private company more practical influence over public debate than many ministries once had.
The citizen is therefore caught between two powerful systems.
One system has police, law, tax notices, licensing powers and prisons.The other has algorithms, terms of service, moderation teams, recommendation engines, ad markets and control over visibility.
Both claim to protect the public. Both can also silence the public.
This is the new crisis of freedom of speech: expression is no longer threatened only by censorship; it is also shaped by platform architecture, opaque moderation and algorithmic power.
Free speech is not a decorative right
Freedom of speech is not simply the right to talk.
It is the right to think aloud in public. It is the right to criticize authority, expose corruption, publish uncomfortable facts, defend unpopular ideas, question social norms, challenge religious or political orthodoxy, demand accountability and participate in collective life.
Article 19 of the Universal Declaration of Human Rights recognizes the right to freedom of opinion and expression, including the freedom to seek, receive and impart information and ideas through any media and regardless of frontiers. The International Covenant on Civil and Political Rights similarly protects the right to hold opinions and express information and ideas of all kinds, while allowing only limited restrictions under defined conditions.
This matters because freedom of speech is not only an individual liberty. It is democratic infrastructure.
Without speech, elections become theatre.Without speech, corruption hides.Without speech, courts lack public scrutiny.Without speech, minorities cannot defend themselves.Without speech, citizens become subjects.
Every authoritarian system understands this. That is why it attacks speech early.
The old censor has not disappeared
Digital platforms may be new, but state censorship remains brutally alive.
Governments still arrest journalists, ban media outlets, intimidate editors, prosecute activists, block websites and create laws so vague that citizens begin censoring themselves. The methods vary. Some states use sedition, defamation, anti-terror laws, cybercrime statutes, public order rules, national security laws or emergency powers. Others use tax raids, advertising pressure, ownership capture, licence cancellations or online harassment by ruling-party ecosystems.
The result is often the same: the citizen learns that speech has consequences.
Reporters Without Borders’ 2026 World Press Freedom Index said global press freedom was at a 25-year low and noted that the legal indicator had deteriorated in more than 60 percent of countries assessed, including India, Egypt, Israel and Georgia. RSF described the criminalization of journalism through emergency legislation, common law misuse and other legal pressures as a global phenomenon.
That is an important warning. The decline of speech rarely begins with a formal announcement that freedom is over. It begins with selective prosecutions, vague laws and examples made of a few people so that thousands learn to remain silent.
The most effective censorship is not the censorship that bans every word. It is the censorship that makes citizens calculate risk before speaking.
Press freedom is the frontline of public speech
Journalists are often the first target because journalism converts private wrongdoing into public knowledge.
A government can survive opposition slogans. It fears evidence. A corrupt system fears documents, investigations, leaked files, field reporting, casualty counts, eyewitness testimony and financial trails. That is why independent media is so often attacked.
RSF’s India country profile says journalists critical of the government face online harassment, intimidation, threats, physical attacks, criminal prosecutions and arbitrary arrests. It also describes India as one of the world’s more dangerous countries for media professionals, with an average of two to three journalists killed due to their work every year.
India is not alone. Hong Kong’s press freedom environment has deteriorated sharply since the 2019 protests, with major outlets shut and journalists facing legal risks. A recent AP report noted that Hong Kong now ranks 140th in the World Press Freedom Index, while cases involving journalists and former media leaders continue to raise concerns about shrinking space for independent reporting.
The pattern is global. Where power becomes insecure, journalism becomes dangerous.
The state now uses safety language to regulate speech
The most common modern justification for speech restriction is not ideology. It is safety.
Governments say they are fighting misinformation, hate speech, terrorism, cybercrime, child abuse material, communal violence, foreign interference, election manipulation or national security threats. Many of these concerns are real. A democratic society cannot ignore incitement, coordinated disinformation, targeted harassment, child exploitation or violence planned through digital networks.
But safety language can become a censorship machine when legal definitions are vague and executive power is unchecked.
A law against misinformation may become a law against criticism.A law against hate may become a law against dissent.A law against foreign interference may become a law against independent civil society.A law against terrorism may become a law against uncomfortable reporting.
The danger is not that governments regulate speech. Every constitutional system allows some limits. The danger is that governments regulate speech without transparency, necessity, proportionality and independent review.
The UN Special Rapporteur on freedom of expression warned in a 2025 report on elections in the digital age that declining freedom of expression must be reversed if public trust in elections is to be restored. That warning matters because misinformation cannot be fought by destroying the expressive environment that makes democracy credible.
Democracies must fight falsehoods without creating ministries of truth.
India’s fact-check unit debate shows the dilemma
India’s debate over government fact-checking shows how difficult the speech question has become.
The problem of misinformation is serious. False claims about health, elections, communal violence, disasters and government schemes can cause real damage. A state has a legitimate interest in correcting false information and protecting citizens from harmful deception.
But the question is: who decides what is false when the speech concerns the government itself?
The 2023 amendment to India’s IT Rules sought to create a government fact-check unit with power to identify online content about the central government as fake, false or misleading. Critics argued that allowing the government to be judge in its own cause would chill political criticism and pressure intermediaries into removing lawful speech. The Bombay High Court ultimately struck down the amendment in September 2024 as unconstitutional, according to Internet Freedom Foundation’s legal summary.
The case is important because it captures the central tension of modern speech regulation. Misinformation is real. But state-controlled truth can be more dangerous than misinformation if it is used to protect power from scrutiny.
A democratic government can correct falsehood. It should not control truth.
Platforms are not neutral public squares
The second major pressure on speech comes from platforms.
Social media companies often describe themselves as open spaces for expression. In reality, they are governed environments. Their algorithms decide visibility. Their rules define acceptable speech. Their moderation systems remove or label content. Their advertising models reward attention. Their recommendation engines shape what citizens see first, often, and repeatedly.
This makes platforms more than private businesses. They are public communication infrastructure operated by private companies.
A person may technically be free to speak, but if the algorithm buries the post, the practical effect is silence. A journalist may publish a story, but if a platform demonetizes it, labels it, downranks it or suspends the account, the reach changes dramatically. A political movement may exist, but if platform rules or government pressure remove its accounts, its public visibility can collapse.
This is not the same as state censorship. Platforms are not governments. They have the right and duty to moderate illegal and harmful content. But their scale gives them quasi-public power.
The modern citizen’s speech is therefore shaped not only by law, but by code.
Moderation is necessary, but opacity is dangerous
A platform without moderation becomes unusable. Spam, abuse, scams, child exploitation, violent threats, targeted harassment, terrorist content and manipulated media can destroy the online public sphere. The question is not whether platforms should moderate. They must.
The question is how.
If moderation is opaque, users do not know why content was removed. If appeals are weak, mistakes cannot be corrected. If rules are vague, enforcement becomes arbitrary. If algorithms reward outrage, platforms profit from the very toxicity they later claim to control. If governments send secret takedown requests, users may never know whether content disappeared because it broke platform rules or because the state demanded removal.
Meta’s transparency reporting on India says it restricted access to more than 16,000 items in India through law enforcement takedown request channels for alleged violations of local regulations. Meta says local-law restrictions are applied only in the country or region where the content is alleged to be illegal.
X’s transparency page for India lists regular Information Technology Rules reports, showing how platform-state compliance has become a continuous administrative process rather than an exceptional event.
These transparency systems are useful, but they also reveal the scale of the issue. Speech today is often governed through a triangular relationship: state requests, platform compliance and user uncertainty.
The platform-state alliance can become invisible censorship
The most dangerous form of modern censorship may not be open banning. It may be indirect pressure.
A government may not prosecute a citizen directly. Instead, it may pressure platforms to remove content. A platform may comply to avoid legal liability, raids, fines, licence issues, criminal exposure or market access problems. The user receives a notice, or sometimes no meaningful explanation at all. The state avoids direct accountability. The platform avoids confrontation. The citizen loses speech.
This is censorship by outsourcing.
In 2025, the Committee to Protect Journalists reported that X accused India of press censorship after blocking news outlets’ accounts, and noted that X had previously expressed concern over Indian government demands to block more than 8,000 accounts. CPJ said X had asked for executive blocking orders to be made public.
The issue is not whether every blocked account was lawful or unlawful. The issue is whether speech restrictions affecting public debate are transparent, reasoned, reviewable and proportionate.
When takedown orders are secret, democracy suffers. Citizens cannot debate what they cannot see.
Europe’s Digital Services Act shows the regulatory turn
The European Union has taken a different approach through the Digital Services Act. The DSA aims to make the online environment safer and more trustworthy by imposing obligations on digital services, especially large platforms and search engines. The European Commission describes it as a framework for transparency, accountability and user protection online.
The DSA is important because it recognizes that platforms cannot be left entirely to self-regulate. It demands clearer processes, risk assessments, transparency measures and stronger accountability from very large online platforms.
But even well-intentioned regulation carries risks.
If platform regulation becomes too aggressive, platforms may over-remove content to avoid liability. If states define harm too broadly, lawful speech may be suppressed. If European standards are applied globally by platforms for simplicity, rules designed for one legal culture may affect speech elsewhere. Policy analysts have warned that some DSA transparency systems, such as statements of reasons for content moderation decisions, are useful but still raise questions about how EU content rules may influence global platform behaviour.
The lesson is clear: platform accountability is necessary, but regulation must not become bureaucratic censorship.
The algorithm is now an editor without a name
Traditional censorship removes speech. Algorithmic control can do something more subtle: it controls attention.
On social media, the core question is not only whether speech is allowed. It is whether speech is seen.
Algorithms decide which posts appear first, which videos autoplay, which comments are hidden, which creators grow, which topics trend and which emotional triggers receive more reach. These decisions are not neutral. They are shaped by business incentives, engagement metrics, safety systems, advertiser preferences and sometimes political pressure.
This creates a new kind of speech inequality.
The loudest voice may not be the most truthful.The most viral claim may not be the most important.The most visible outrage may not represent public opinion.The most amplified creator may not be the most credible.
In earlier times, editors in newspapers made visible decisions. Readers could identify the institution shaping the front page. Today, the “editor” is often an algorithmic system no ordinary citizen can inspect.
This is why freedom of speech must now include freedom from opaque manipulation of public attention.
Misinformation has made the free speech debate harder
The free speech debate became more complicated because misinformation is not imaginary.
Falsehoods can cause violence, undermine elections, spread medical danger, damage reputations, incite mobs, exploit disasters and polarize society. During elections, digital disinformation can erode public trust. During conflicts, false claims can inflame hatred. During pandemics, false medical advice can kill.
Freedom House’s Freedom on the Net 2025 report found that global internet freedom declined for the 15th consecutive year and warned that the internet is more controlled and manipulated than ever before, with governments increasingly deploying censorship, surveillance and control over emerging technologies.
This creates a difficult policy problem. The same digital environment that enables dissent also enables manipulation. The same encrypted app that protects activists can spread rumours. The same platform that exposes injustice can amplify hate. The same viral video that documents abuse can also be fake.
The correct answer is not absolute laissez-faire. But neither is state-controlled information.
The answer must be layered: independent fact-checking, media literacy, transparent platform rules, accountable moderation, public-interest journalism, civic education, rapid correction systems, and narrow legal action against direct incitement or coordinated harmful deception.
A society that cannot distinguish between disagreement and danger will not protect speech for long.
Hate speech is the hardest boundary
Every free society struggles with the line between offensive speech and dangerous speech.
Democracy must tolerate criticism, satire, blasphemy, ideological disagreement, harsh political language and unpopular opinion. If speech is protected only when polite, it is not free speech. But democracy cannot ignore speech that directly incites violence, dehumanizes communities, organizes targeted harassment or creates imminent danger.
This line is difficult because states often exploit it.
Authoritarian governments call criticism hate. Majoritarian movements call minority assertion provocative. Platforms may remove controversial but lawful speech while failing to act against coordinated abuse. Activists may demand removal of speech they find offensive, while governments use those demands to justify wider control.
The principle should be clear: speech restrictions must target specific harm, not discomfort. They must be precise, lawful, necessary, proportionate and independently reviewable.
A society cannot protect dignity by destroying liberty. But it also cannot protect liberty by ignoring organized dehumanization.
The journalist, the activist and the ordinary citizen now face different risks
Freedom of speech does not operate equally for everyone.
A powerful politician can claim censorship while commanding television time, party machinery and millions of followers. A journalist in a small town faces defamation cases, police pressure, physical threats and loss of livelihood. A woman speaking online faces abuse and sexualized threats. A minority citizen may face mob targeting. A comedian may face criminal complaints across states. A student may face disciplinary action. A whistle-blower may lose employment.
This is why the formal right to speak is not enough.
The real question is whether citizens can speak without disproportionate fear.
If only the powerful can speak freely, the system has speech privilege, not speech freedom.
India’s constitutional balance is being tested
India’s Constitution protects freedom of speech and expression under Article 19(1)(a), subject to reasonable restrictions under Article 19(2). This framework recognizes both liberty and social order. India is not an absolutist free speech system like the American First Amendment model. It allows restrictions for sovereignty, security, public order, decency, morality, contempt of court, defamation, incitement and related grounds.
The challenge is not the existence of restrictions. The challenge is their misuse.
In a vast and diverse country, some restrictions are necessary. India has communal tensions, terrorism concerns, caste violence, misinformation problems, border conflicts and deep social sensitivities. But a democracy of India’s scale must be especially careful because the state’s coercive power is enormous.
A legal notice from the state is not a small inconvenience for an ordinary citizen. It can mean years of litigation, police visits, reputational damage, job loss and fear.
India therefore needs a speech framework that is firm against incitement and violence, but protective of criticism, satire, journalism, political dissent and academic debate.
The more powerful the state becomes, the more important speech protection becomes.
Free speech is also under pressure from society
The state and platforms are not the only actors pressuring speech.
Society itself can become censorious. Political supporters attack critics. Religious groups demand bans. Corporate advertisers withdraw support. Online mobs threaten creators. Ideological communities demand conformity. Universities avoid controversial speakers. Publishers fear backlash. Artists self-censor. Ordinary people avoid expressing opinions because every statement can be clipped, distorted, archived and weaponized.
This is social censorship.
It may not involve police, but it creates fear. The punishment is not prison; it is harassment, boycott, cancellation, violence, job loss or reputational destruction.
A free society needs more than legal speech rights. It needs a culture of tolerance for disagreement.
This does not mean freedom from criticism. Speech has consequences. But there is a difference between criticism and coercion, between accountability and intimidation, between disagreement and destruction.
Democracy requires citizens capable of hearing things they dislike.
The business model of outrage weakens speech quality
Platforms often say they protect expression. But their business model rewards attention, not truth.
Anger travels faster than nuance. Insult travels faster than analysis. Conspiracy travels faster than correction. Emotion keeps users engaged. Engagement produces data. Data produces advertising revenue.
This does not mean platforms deliberately want bad speech. But their incentives can produce a public sphere in which extreme, emotional and divisive content receives more visibility than careful argument.
The result is paradoxical: there may be more speech than ever, but less meaningful deliberation.
Citizens are not censored; they are flooded.Truth is not banned; it is buried.Debate is not eliminated; it is degraded.
Freedom of speech is weakened not only when voices are silenced, but also when the information environment becomes so polluted that responsible speech cannot compete.
AI will intensify the crisis
Artificial intelligence will make the speech debate even harder.
AI can generate fake videos, synthetic voices, automated comments, mass propaganda, impersonated journalists, forged documents and personalized political messages. It can also help detect hate speech, translate content, moderate at scale and improve accessibility.
The technology is not inherently anti-speech. But it changes the scale of manipulation.
In the AI age, the problem will not only be whether a person can speak. It will be whether citizens can know who is speaking.
Is this a real voter or a bot?Is this a journalist or an impersonation?Is this a video or a deepfake?Is this public opinion or synthetic amplification?Is this criticism or coordinated manipulation?
If authenticity collapses, trust collapses. And when trust collapses, governments often demand more control.
AI therefore creates a two-sided threat: more manipulation by bad actors and more censorship justified by fear of manipulation.
The democratic answer must be transparency, provenance tools, media literacy, platform accountability and narrow regulation — not panic-based speech control.
The answer is not absolutism
Some free speech defenders argue that almost all speech should be left alone. That position has moral appeal because censorship powers are often abused. But absolute speech freedom can ignore real harms: incitement, targeted harassment, child exploitation, fraud, doxxing, threats and coordinated manipulation.
On the other side, some regulation advocates treat speech as something that must be managed for social harmony. That position can quickly become authoritarian because power decides what harmony means.
A serious democracy must reject both extremes.
Speech must be protected broadly. Restrictions must be narrow. Harm must be defined precisely. The state must carry the burden of justification. Platforms must provide due process. Courts must act as safeguards. Citizens must develop resilience against offensive and false speech without demanding bans for every discomfort.
Free speech is not the absence of rules. It is the presence of rules that prevent power from silencing legitimate expression.
What platforms must do
Platforms cannot continue hiding behind neutrality.
They must publish clear moderation rules in accessible language.They must notify users when content is removed or restricted.They must provide meaningful appeals.They must disclose government takedown demands wherever legally possible.They must label state-affiliated media and coordinated influence operations.They must reduce opaque algorithmic amplification of harmful manipulation.They must protect journalists, activists and vulnerable groups from coordinated harassment.They must support independent research access while protecting privacy.They must be transparent about political advertising and paid influence.
This is not censorship. This is accountability.
Private companies now govern public speech at a scale never before seen. Their power must be matched by responsibility.
What states must do
States must also accept limits.
They should publish takedown orders unless a narrow and justified exception applies.They should ensure independent judicial or quasi-judicial review of speech restrictions.They should define misinformation laws narrowly.They should not become arbiters of truth about themselves.They should protect journalists from violence and legal harassment.They should reform criminal defamation and vague public order provisions.They should avoid internet shutdowns except in truly exceptional circumstances.They should protect encryption and privacy instead of treating anonymity as criminal by default.They should ensure that national security is not used as a blanket argument against scrutiny.
A state that cannot tolerate criticism is not protecting democracy. It is protecting power.
What citizens must do
Citizens also carry responsibility.
Freedom of speech does not mean freedom from judgment. It does not mean forwarding rumours. It does not mean threatening opponents. It does not mean abusing women, minorities, journalists or critics. It does not mean turning public debate into mob punishment.
A healthy speech culture requires habits: reading before reacting, verifying before sharing, criticizing without dehumanizing, defending the speech rights of opponents, and resisting the temptation to demand censorship whenever speech becomes uncomfortable.
The citizen is not only a rights-holder. The citizen is also a participant in the public sphere.
If citizens use speech only as a weapon, they should not be surprised when the state arrives claiming to restore order.
The final test
Governments fear it because it exposes authority.Platforms monetize it because it captures attention.Political movements manipulate it because it mobilizes emotion.Citizens need it because it is the foundation of dignity and democracy.
This is why the struggle over speech will define the next phase of politics.
The old question was: can the state censor the citizen?The new question is larger: who controls the conditions under which citizens speak, hear, trust, disagree and organize?
A free society cannot allow governments to silence dissent in the name of safety.It cannot allow platforms to invisibly govern public debate without accountability.It cannot allow mobs to intimidate citizens into silence.It cannot allow misinformation to destroy trust.It cannot allow regulation to become censorship.
The balance is difficult, but the principle is simple: power must not control truth.
Freedom of speech will survive only if it is defended against both public coercion and private opacity. It must be protected from the jailer and the algorithm, from the censor and the mob, from the minister and the monopoly.
Because once speech is controlled, every other right becomes dependent on permission.
And a democracy where citizens speak only by permission is no longer a democracy.
