From the perspective of 2026, the decade-long evolution of post-truth politics and online falsehoods appears less like a sudden phenomenon of the mid-2010s and more like the beginning of a lasting process of truth decay—from social media’s amplification of misinformation to the systematic production of fabricated realities by AI, writes Yury Kolotaev.
The year 2016 marked a significant milestone at the intersection of digital technology and global politics. Social media had already influenced elections and social conflicts for some time. However, 2016 saw the systematic rise of digital misinformation and disinformation. One notable aspect of this development was the widespread claim that societies had entered a “post-truth” era. The term “post-truth” (alongside “fake news”) was named “Word of the Year” by several institutions. A decade later, post-truth is an enduring feature of a weaponised digital environment.
Post-truth refers to political communication disconnected from verified information, relying on emotional appeals and collective sentiment. Events in the late 2010s, including Brexit, the Myanmar crisis, the Cambridge Analytica scandal, and Europe’s right-wing movements, highlighted the technological preconditions of post-truth politics. While many had initially viewed the internet as a tool for open, interconnected access to information, algorithm-driven social media instead accelerated an erosion of social trust.
Little improvement has occurred in addressing these social challenges; instead, societies have become increasingly fragmented. Like social media before it, artificial intelligence offers powerful tools to actors with harmful intentions. Whereas social media spread misinformation, AI generates content, often termed “AI slop”, that distorts public discourse.
Developments in the United States are noteworthy. President Donald Trump pioneered assertive social media use, normalising “alternative facts”. Despite moderation and deplatforming, he was re-elected in 2024. Throughout 2025, the Trump administration shared AI-generated videos and memes, using political humour to blur reality and fiction. For example, Trump reposted a video showing a plane dumping waste on “No Kings” protesters. Similar content is appearing ahead of the 2026 midterms. This trend normalised AI political imagery, with media reporting a sharp increase in synthetic visuals.
Low-quality AI content has proliferated in 2025–2026 conflicts and elections, fuelling post-truth narratives across Africa, Asia, and the Middle East. Recycled footage has intensified disinformation. Expanding digital ecosystems have faced pressure from external and domestic actors, distorting events and outpacing verification.
The key mid-2020s shift is from election disinformation to wartime information threats: fabricated missile attacks, repurposed pre-war AI art, and downed aircraft amplify panic and misrepresent events. AI-slop appears within an hour of strikes. As a result, AI-driven psychological operations are becoming a new military standard that amplifies emotion over accuracy during military clashes (e.g., the India–Pakistan dispute and the Thailand–Cambodia border crisis).
AI development should not be framed as catastrophic. Criticising technology—as with earlier debates over social media—has proven limited in fostering trust. Repeated efforts to address online disinformation have often intensified polarisation, blurring the line between protection and censorship. This happened during the COVID-19 pandemic, resulting in a post-truth “infodemic”. In 2020-2021 even crucial health information became politicised, leading to emotionally charged debates in which neither the WHO nor national authorities were able to establish a broad consensus.
Synthetic media provide even fewer straightforward answers. AI-generated texts and videos are not a faster version of disinformation but a qualitative shift. The 2020 infodemic shows that framing suspicious information as a security problem erodes trust in both unofficial and official sources. Similarly, when citizens suspect that all media may be AI-manufactured, social consensus begins to crumble. Fortunately, there are various self-correction mechanisms.
In an attention-driven digital environment, the problem is not only misleading content. The deeper issue is structural incentives that reward emotional engagement over accuracy. Since information has a limited life cycle, it becomes trivial as public attention shifts. Thus, the most effective defence is personal awareness that content failing to provoke engagement quickly becomes obsolete.
More complex post-truth issues arise from personalised search engines and AI chatbots. The pursuit of rapid access to highly plausible, but not necessarily verified, information creates significant vulnerabilities. The issue is no longer merely social media feeds and content, but primary sources of information themselves.
This represents a shift in “media power” from social networks to new personalised AI platforms. In 2025-2026, this prompted a wave of state-led initiatives to develop national or regional large language models (LLMs) in China, South Korea, Russia, and the UAE, designed to reflect specific cultural contexts and security priorities. This marks a new front line of the post-truth struggle: not just fighting falsehoods but ensuring that the primary sources of information themselves do not become tools of foreign political influence.
Post-truth conditions push political actors beyond content moderation to platform and infrastructure control. The US Federal agencies attempt to prioritise governmental sources in chatbot responses, the European Union’s expanded Digital Services Act and AI Act now strive to cover generative AI with transparency obligations, while the BRICS+ nations are exploring cross-verification mechanisms for synthetic media. The competition is no longer just about narratives, but AI models, content identification and prompt results.
The discussion still centres on digital platforms in politics. Major AI companies (OpenAI, Google, Anthropic) may hold information monopolies, but competition and platform pluralism can counteract this. Social network monopolies once seemed unbreakable. Now, regional and national AI alternatives (LatamGPT, Sarvam AI) emerge alongside decentralised networks, offering options. Ultimately, with AI or social networks, the key is balancing open information flow with responsible social conduct.
One lesson from AI development is that technology amplifies, but does not create, political and informational fractures. The battle against post-truth is not just about technological fixes or individual critical thinking. It is a matter of political will, institutional capacity, and social consensus. It also represents the constant struggle for attention and emotions.
The past decade of post-truth has demonstrated that simply letting information circulate freely is no longer practical. Rather than choosing between regulation and freedom, societies face different models of governance. These include platform-led moderation seen in the West, calls for fair AI oversight in the Global South, and unified national AI systems in other regions. In the coming decade, states and communities are likely to reassert control over digital spaces. Without governance adapted to local contexts, post-truth dynamics will not just persist but deepen, eroding inter-state communication and trust.
