The digital age has fundamentally altered the landscape of truth. As Mark Twain famously noted, a lie can cross the globe while the truth is still tying its laces, but today’s infrastructure—driven by social media algorithms—has turned that phenomenon into a systemic crisis. While we often debate whether the state should combat misinformation, we must be careful. A broad, government-led mandate to police “truth” is inherently dangerous; it threatens the pluralism and freedom of speech essential to democratic society. When states attempt to define what is true, they risk creating tools for censorship that can be easily abused to silence dissent. Therefore, the state’s duty shouldn’t be to act as an arbiter of ideas, but rather to protect citizens from foreseeable, rights-harming violence orchestrated by the unchecked architecture of virality.
To avoid the slippery slope of state-sponsored censorship, we need a precise, narrow standard for when intervention is necessary. This duty should not be triggered by the mere existence of a falsehood, but by a specific confluence of four factors: the information must be demonstrably false; it must be amplified at a massive scale; it must pose an imminent, objective risk to public safety or human rights; and any government response must be proportionate, targeted, and time-limited. By shifting the focus from the content of the message to its engineered reach, we protect the “marketplace of ideas” while acknowledging that some online explosions—such as those that incite racially motivated violence or public disorder—are not organic expressions but manufactured consequences of platform design and recommender systems.
Current legal frameworks, such as the UK’s Online Safety Act, are fundamentally ill-equipped to address this reality. The Act focuses on individual criminal intent, which requires proving that a specific user intended to cause harm. While this is useful for individual prosecutions, it fails to address the “lawful-but-harmful” misinformation that spreads like wildfire through recommender algorithms. Business models that prioritize engagement often reward sensationalist lies, creating a structural problem that the current law ignores. As seen during the 2024 unrest, misinformation can incite real-world destruction in a matter of hours, far faster than the legal system or traditional media-literacy campaigns can respond. We cannot rely on “educational” initiatives to solve an emergency that is unfolding in real time; we need structural interventions that operate at the speed of the crisis.
The template for this transition already exists within the European Union’s Digital Services Act. This framework moves away from content-policing and toward a “risk-management” model. Instead of having the government judge the truth, the law requires platforms to assess and mitigate the systemic risks inherent in their own architecture. This aligns with international human rights standards because it focuses on the mechanism of harm—algorithmic amplification—rather than the silencing of political or heterodox views. By forcing platforms to be transparent about how their systems work and by requiring them to introduce “friction” into the sharing process, we can curb the velocity of dangerous lies without ever needing a government-run “Ministry of Truth.”
The goal of this “epistemic fire brigade” is to introduce speed-bumps into the machinery of viral misinformation. When an event crosses the threshold into foreseeable violence, regulators could mandate temporary downranking, the promotion of verified public-interest corrections, or the introduction of friction in the resharing process. Most importantly, these actions must be procedural, transparent, and subject to immediate judicial appeal. We are not calling for the destruction of free expression; we are advocating for the regulation of the conduits that give that expression an unnatural, artificial intensity. Friction, unlike prohibition, preserves the open exchange of ideas while stripping dangerous falsehoods of their engineered, destructive velocity.
In conclusion, the state’s obligation must be defined by its limits. A government that tries to decide what is true will eventually destroy the very liberty it claims to protect. However, a state that remains a passive bystander while private companies monetize and accelerate the path to real-world violence is equally failing its citizens. Our duty is not to protect society from error, but to protect the public from the algorithmic amplification of lies that lead, foreseeably and directly, to the violation of fundamental rights. By targeting the architecture of the digital world rather than the speaker, we can build a safer society that remains committed to the principle of free, independent, and autonomous citizenship.