How Tech Laws Are Redefining Free Speech Online

1 August 2026

The internet was supposed to be the ultimate free speech platform. A global town square where anyone could speak, publish, and debate without permission. That vision, born in the 1990s, is now colliding with a wave of new laws designed to regulate what we say, how platforms moderate content, and who is held accountable for the consequences. The result is not simply a restriction on speech. It is a fundamental redefinition of what free speech means in a digital age where private companies control the infrastructure, algorithms amplify outrage, and governments across the world demand different versions of the truth.

We are no longer asking whether platforms should moderate content. They always have. The real question is who decides the rules, what those rules protect, and how they reshape the public sphere in ways most users do not yet fully grasp.

The End of the Platform Era

For two decades, the dominant legal framework for online speech in the United States was Section 230 of the Communications Decency Act. Passed in 1996, it granted interactive computer services immunity from liability for user-generated content while allowing them to moderate in good faith. This single provision created the modern internet. It allowed YouTube to host millions of videos without facing lawsuits for every copyright claim, defamatory comment, or hateful rant. It let Facebook and Twitter scale to billions of users because they were not treated as publishers.

That era is over. Not because Section 230 has been repealed, but because new laws in Europe, the United Kingdom, and increasingly in US states are layering obligations on top of it that change the calculus entirely. The Digital Services Act in the European Union, the Online Safety Act in the UK, and a patchwork of state-level laws in the US are forcing platforms to take responsibility for content in ways that were previously unthinkable.

The shift is from immunity to accountability, but the accountability is not symmetrical. It is not about protecting individual speakers from punishment. It is about forcing platforms to police speech according to government-defined standards. That is a very different thing.

The DSA and the New European Standard

The Digital Services Act, which applies to any platform serving EU users regardless of where the company is headquartered, represents the most comprehensive attempt to regulate online speech to date. It introduces a tiered system of obligations. Very large online platforms, those with more than 45 million monthly active users in the EU, face the strictest requirements.

The DSA does not tell platforms what speech is legal. Instead, it requires them to implement systemic risk assessments. Platforms must identify risks related to illegal content, disinformation, and fundamental rights. They must then take mitigation measures. This sounds reasonable on paper, but the practical effect is that platforms must build moderation systems that err on the side of removal.

Here is the trade-off that most commentators miss. The DSA creates strong incentives to over-remove rather than under-remove. If a platform leaves illegal content up, it faces fines of up to 6 percent of global turnover. If it removes legal content that merely resembles illegal content, the penalty is much lower. A rational platform will therefore train its algorithms to be aggressive. That means legitimate speech gets caught in the net.

Consider the example of a satire account that mimics the style of a terrorist group's propaganda. Under the DSA, the platform cannot easily distinguish between satire and genuine threat. The cost of getting it wrong is enormous. The cost of removing the satire account is just a few complaints. The platform removes it. Free speech loses.

The DSA also introduces a complaint mechanism, but it is slow and bureaucratic. Users can appeal moderation decisions, but the appeal process takes weeks. In the meantime, the content stays down. For time-sensitive speech, such as political commentary during an election, that delay is effectively a silencing.

The UK Online Safety Act and the Age Verification Problem

The UK's Online Safety Act takes a different approach. Instead of focusing on all speech, it creates specific duties of care for platforms regarding harmful content. The act applies to user-to-user services and search engines, requiring them to remove illegal content quickly and to protect children from legal but harmful material.

The age verification provisions are the most contentious. Platforms that host adult content or that allow user-generated content likely to be accessed by children must implement robust age checks. The practical result is that many platforms now ask for government-issued ID or facial age estimation. That chills anonymous speech, which has always been a cornerstone of online expression. People who want to discuss sensitive topics, such as domestic abuse, sexual identity, or political dissent, may not want to provide a passport to a private company.

The act also requires platforms to assess the risk that their services will be used to commit priority offenses, which include things like hate crimes, harassment, and the encouragement of self-harm. This creates a duty to monitor for signs of such behavior, which pushes platforms toward proactive scanning of private messages. End-to-end encryption is directly threatened by this provision. If a platform cannot see the content, it cannot assess the risk. The UK government has repeatedly said it will not ban encryption, but the act's requirements make it nearly impossible to offer encrypted messaging without some form of client-side scanning.

The deeper issue is that the Online Safety Act, like the DSA, outsources the definition of harm to platforms. The act lists categories of harmful content, but it does not define them precisely. Platforms must use their own judgment, which leads to inconsistent enforcement. What is considered harmful by a platform's moderation team in London may not be harmful in Mumbai or Austin.

The US Patchwork and the Return of State Power

In the United States, the federal government has not passed a comprehensive online speech law since 1996. But the states are filling the void. Florida and Texas have passed laws that restrict platforms' ability to moderate content based on viewpoint. These laws are currently tied up in litigation, but they signal a major shift in the political landscape.

The Florida law, SB 7072, prohibits platforms from banning political candidates and requires them to publish their moderation standards. The Texas law, HB 20, goes further by prohibiting platforms from removing content based on the viewpoint expressed. Both laws are premised on the idea that platforms are common carriers, like telephone companies, and should not discriminate against any speaker.

This is a radically different vision from the European approach. The EU is telling platforms to remove more content. Texas is telling them to remove less. The result is that a platform like X (formerly Twitter) must now behave differently in Texas than in Germany. It must keep hate speech up in Texas while removing it in Germany. This creates a technical nightmare and a legal minefield.

The Supreme Court heard arguments in these cases in 2024, and the outcome will shape the future of online speech in the US. The key question is whether platforms are more like newspapers or more like common carriers. If they are newspapers, they have a First Amendment right to decide what content to publish. If they are common carriers, they cannot discriminate. The answer likely lies in between, but the Court's decision will determine whether the internet remains a curated experience or becomes a raw transmission medium.

The Algorithmic Amplification Problem

All of these laws ignore the elephant in the room: algorithms. The speech that users see is not a neutral reflection of what others post. It is a highly personalized feed selected by machine learning systems optimized for engagement. The legal frameworks treat platforms as if they were simply hosting content that users choose to view. In reality, platforms actively shape what users see, how often they see it, and in what context.

The DSA addresses this by requiring transparency about recommendation systems and giving users the option to see content without algorithmic ranking. But this is a weak remedy. Most users will not change their default settings. They will continue to scroll through algorithmically curated feeds, and the algorithm will continue to amplify outrage, conflict, and emotionally charged content because that is what drives engagement.

The deeper problem is that algorithms are not speech. They are a form of behavior modification. When a platform decides to promote a particular video to millions of users, it is not merely publishing that speech. It is actively distributing it. The law has not yet developed a coherent framework for treating algorithmic amplification as a distinct form of expression or regulation.

Some scholars argue that algorithms should be protected as speech because they encode a point of view. Others argue that they are more like infrastructure and should be subject to neutrality requirements. The truth is that neither approach fully captures the reality. Algorithms are a hybrid of editorial judgment and automated decision-making. They are not human, but they are not neutral either.

Practical Implications for Users and Creators

What does this mean for the average person who wants to speak online? It means that speech is no longer governed by a single set of rules. It is governed by a complex web of jurisdiction-specific laws, platform policies, and algorithmic systems that are largely opaque.

If you are a content creator, you need to understand that your content may be legal in your home country but illegal in the EU. The DSA applies to any platform that serves EU users, so if your content is visible in Europe, it must comply with EU rules. That means you may be subject to removal, not because your content is illegal in your jurisdiction, but because it violates a standard set in Brussels.

If you are a journalist, you face a different problem. The new laws require platforms to remove illegal content quickly. This creates pressure to remove content that is merely controversial or that has not yet been verified. The old model of "publish first, correct later" is becoming untenable. Platforms are increasingly requiring evidence of legality before allowing publication, which chills investigative reporting.

If you are an activist, you need to be aware that anonymous speech is under threat. Both the UK and the EU are pushing for greater identity verification. This is not necessarily a bad thing if it reduces harassment and abuse. But it also means that governments can identify dissidents, and that platforms can be compelled to hand over user data.

The Misconception About Moderation

A common misconception is that content moderation is a simple binary: either a post stays up or it comes down. The reality is far more complex. Platforms use a range of measures, including downranking, demonetization, shadow banning, and labeling. These measures do not remove speech, but they reduce its visibility. They are the preferred tools of modern platforms because they are less legally risky than outright removal.

Downranking is particularly insidious. The content stays up, so the platform can argue it is not censoring. But the algorithm ensures that almost no one sees it. This is a form of de facto censorship that is nearly impossible to challenge. There is no transparency requirement that forces platforms to explain why a particular post has been downranked. The user sees a drop in engagement but has no way to know what caused it.

The new laws do not address downranking. The DSA requires transparency about recommendation systems, but it does not require platforms to explain individual ranking decisions. This is a significant gap. A platform can comply with the letter of the law while still silencing speakers through algorithmic invisibility.

The Trade-Off Between Safety and Freedom

There is no perfect system. Every approach to online speech involves trade-offs. The European model prioritizes safety and social cohesion over individual expression. The American model, at least in its traditional form, prioritizes individual expression over collective safety. The new state laws in the US are trying to shift the balance toward a more laissez-faire approach, but they are doing so in a way that may be unconstitutional.

The key insight is that free speech online is not a natural state. It is a product of specific legal and technical choices. Section 230 created the conditions for open platforms. The DSA and the Online Safety Act are now changing those conditions. The result will not be the end of free speech, but it will be a different kind of free speech, one that is more regulated, more national, and more dependent on the whims of private companies.

What Should Be Done

The most important step is to push for greater transparency in algorithmic decision-making. Users should have the right to know why they are seeing certain content and why their content is being suppressed. This does not require revealing proprietary algorithms. It requires explaining the factors that led to a specific decision in a specific case.

The second step is to harmonize legal standards across jurisdictions. The current patchwork of laws is unworkable. A platform cannot simultaneously comply with the DSA, the Online Safety Act, and the Texas law. The result is that platforms will choose to comply with the strictest standards, which means the EU rules will effectively govern the global internet. This is not necessarily bad, but it should be a conscious choice, not an accident.

The third step is to recognize that platforms are not neutral conduits. They are powerful actors with their own interests. The law should treat them accordingly. This means applying antitrust remedies to break up the largest platforms, and it means creating new forms of accountability that do not rely solely on the threat of fines.

The Future of Speech

The next decade will determine whether the internet remains a space for open, diverse, and challenging speech or whether it becomes a highly regulated medium dominated by a few large platforms and a few powerful governments. The technical capacity for free speech has never been greater. Every person with a smartphone can reach a global audience. But the legal and economic structures that enable that capacity are being redesigned in ways that may ultimately limit it.

We need to move beyond the simple narrative of censorship versus freedom. The new laws are not designed to silence dissent in the traditional sense. They are designed to manage risk, to protect children, to reduce polarization, and to hold platforms accountable. But the cumulative effect is a chilling of speech that is difficult to measure and even more difficult to reverse.

The best approach is not to fight every regulation but to engage with it critically. Understand what the laws actually require, what they prohibit, and what they leave to platform discretion. Advocate for rules that are clear, that do not create perverse incentives for over-removal, and that protect the rights of speakers as much as they protect the safety of audiences.

Free speech online is not dying. It is being redefined. The question is whether we will have a voice in that redefinition or whether we will simply adapt to whatever system emerges. The answer depends on whether we are willing to engage with the messy, complex, and often contradictory reality of how speech is actually governed on the internet.