Disinformation — the deliberate production and spread of false or misleading content for political ends — is not new. The Romans did it; Octavian smeared Mark Antony with forged wills before Actium. The Bolsheviks did it. The Cold War CIA and KGB ran it as routine craft — the KGB's Operation INFEKTION spent the 1980s seeding the lie that the U.S. had engineered HIV at Fort Detrick. What is new, in the 2010s and 2020s, is the industrialization of the practice by state actors riding the same algorithmic distribution systems that organize the rest of human communication. Russian operations against the 2016 American election, Chinese operations on Taiwan, Iranian operations on European politics — these are not anomalies or scandals. They are standing tools of state policy, budgeted and staffed like any other arm of statecraft.
The mechanisms exploit features that platforms built for engagement: posts producing strong emotional reactions get amplified; users self-sort into ideologically homogeneous communities; small inputs of carefully crafted content can be pumped by ordinary users into mass distribution. Russia's Internet Research Agency in St. Petersburg ran fake American Black Lives Matter accounts and fake American gun-rights accounts at the same time — even staging dueling rallies at the same Houston mosque in 2016 — with the goal of deepening polarization rather than promoting any single position. The point was never to win an argument; it was to make argument itself feel futile. The newer frontier is generative AI: text, images, audio, and video can now be produced en masse by language models at near-zero marginal cost, in any voice, with no obvious tells — the 2024 robocall mimicking President Biden's voice in New Hampshire was an early, crude preview. The democratic systems most exposed are those with low institutional trust, fragmented media markets, and election cycles that reward novelty over verification — which is to say most of them. The deeper damage is the liar's dividend: once anything can be faked, the powerful can dismiss authentic evidence as fabricated.
Whether democracies can govern themselves under conditions of weaponized information is a genuinely open question. The current generation of safeguards — fact-checking, content moderation, electoral commissions — is being lapped by the technology, and several platforms rolled their moderation back in 2023–2025 just as the tools grew sharper. The next decade will determine whether new institutions emerge to handle this — provenance standards, cryptographic content signing, media-literacy at scale — or whether democracies adapt downwards to a baseline assumption that public information is, on average, weapon-grade, and that nothing seen on a screen can be trusted without proof.