Post-Truth, Deceptive AI, and the Politics of East Asia
May 15, 2026, 12:30 PM
Speaker: Professor Florian Schneider (Director, Leiden Asia Centre, Leiden University)
This public lecture was delivered by Professor Florian Schneider, a scholar of Chinese political communication and digital nationalism. He began by saying that although he normally studies China and its digital technologies, today he wanted to cast the net wider—to a phenomenon that matters to all of us, China or not: how, in a “post-truth” environment flooded with contested information flows, that information becomes politicized. The talk advanced through four factors: humans (the social factor), politics (state actors), political economy (the businesses that built these environments), and the politics of technology itself; it then turned to AI and closed on “what can we do.”
He opened with a familiar case: in 2018, Typhoon Jebi battered the Kansai region and flooded Kansai International Airport, stranding many travelers, including Taiwanese. Rumors spread that Taiwan’s representative office was failing its citizens while China was sending buses to rescue people—and that Taiwanese could board if they said they were Chinese. Social media erupted in anger, and the Taiwanese diplomat in Osaka faced a flood of allegations. It later turned out the information about the buses was false. The case lays bare a tangle of issues, none of them simple: deliberately spread disinformation, misinformation spread for all sorts of reasons, even true information used as a political weapon, and the involvement of international actors. He cited the Spanish sociologist Manuel Castells: communication is always related to power; power relies on the control of communication, while counter-power relies on breaking through it—in network societies, communication power is at the heart of the dynamics.
The Human Factor: Why We Share and Why We Believe
The first factor is people. He gave the example of the “HIV-negative AIDS” panic in Guangdong around 2010: a number of young and middle-aged men developed AIDS-like symptoms, yet doctors uniformly found them HIV-negative; they found one another online, formed a community convinced that “HIV-negative AIDS” was real, and no amount of official or medical correction could persuade them otherwise. The more plausible explanation is psychosomatic—guilt and self-condemnation, in a sexually moralistic society, following extramarital or commercial sexual encounters. Tellingly, it became political: people turned to blaming and distrusting the government for “covering something up,” because the local information environment had long taught citizens not to trust what the Party says—health included.
So why do people share questionable information and believe conspiracies? “Because they’re stupid” is no help, Schneider stressed. Whether we are cognitively declining (microplastics, CO2, shrinking attention spans, falling verbal reasoning) has no scientific consensus, and we are also gaining new capabilities. The key is that people always make trade-offs (outsourcing phone numbers to a device to free attention for something else); “stupidity” is really a social problem—norms, values, and expectations pushing people into shortcuts. More importantly, people share not out of stupidity but because it is pleasurable, funny, because it confirms beliefs they already hold (he admits he does this with left-leaning content), because it restores a sense of agency in an uncontrollable world, and because sharing brings social connection and social capital. When grandma forwards a fake “this prevents cancer” story, she isn’t lying or plotting—she’s doing it because she loves you and wants to show it; misunderstand that, and you respond with heavy-handedness.
This also involves parasocial relationships: we form one-sided “relationships” with celebrities, fictional characters, and communities far too large for anyone to actually know. That is precisely Benedict Anderson’s insight about “imagined communities”—religions and nations alike imagine their connections through shared rituals and symbols (not all Muslims, nor all citizens, actually know one another). The same dynamics apply to fan groups (Taylor Swift fans), who don’t know each other yet can be mobilized as communities. This is the substrate on which everything else is built.
Political Actors: How States Intervene in Information Environments
The second factor is politics. Politicians have always lied, but today’s networked environment carries new and sweeping implications. The starkest example is the dispute over the origins of COVID-19: the US (the Trump administration) pushing the “Wuhan lab leak” theory and terms like “Wuhan virus / China virus” (deeply irresponsible, fueling anti-Asian hatred), and Chinese officials pushing the counter-narrative (Fort Detrick, the US military bringing it in). The result is an environment in which the truth is extremely hard to assess—even though the scientific consensus leans overwhelmingly toward zoonotic transmission at the wet market, a question that should have been debated scientifically was thoroughly politicized.
He noted that the PRC’s, and especially the Party’s, political communication is highly systematic and unapologetic—rooted in a Marxist-Leninist background: the masses are seen as carrying “false consciousness” and as untrustworthy, and only the correctly educated Party leadership can correct that, so “thought work / ideological work” is a moral obligation. Its communication is therefore highly paternalistic (the Party as father and mother teaching the people as children). In this framework, “truth” means “true according to Marxism” (the truth that exposes capitalist exploitation), not necessarily the factual—what matters is whether something is “true,” not whether it is factual. That is why the COVID narrative was quickly “nationalized”: a Beijing art academy reworked the composition of the famous French Revolution painting Liberty Leading the People to depict the fight against the virus as a nationalist story. But a pandemic—like climate change, like the disinformation problem—is fundamentally transnational; pretending COVID is “a Chinese matter” itself creates trouble.
At the covert level there is astroturfing—the “50 Cent Army,” bot networks, and large-scale amplification. Since officials doing it themselves doesn’t land well, they prefer to amplify influencers who seem “more genuine” (including non-Chinese ones). But he cautioned against dismissing all pro-China voices, since some sincerely believe China has positive contributions to make. The real shift is that the goal has moved from “selling a specific narrative” to “generally raising the temperature”—manufacturing distrust and anxiety, because emotional people make bad choices and switch off their cognitive faculties. This is not so different from Kremlin-style disinformation, or operations by various Western actors.
Such operations are of limited use for changing specific outcomes (e.g., the 2016 flooding of Tsai Ing-wen’s Facebook page didn’t meaningfully change politics), but very effective at manufacturing distrust—which can make people slip up at some point. They are also often counterproductive: a Danish broadsheet, for instance, ran a controversial cartoon replacing the stars on China’s flag with virus particles, feeding the “China problem” frame; Chinese netizens fired back with nationalist rhetoric, and Thai netizens borrowed almost identical imagery for “humorous but anti-Chinese” mockery—the whole thing spiraled out of control, hardly to the Party’s liking. Assigning blame is hard: that nationalist wave wasn’t launched by the Party directly but by genuinely angry Chinese netizens, which then drew counterattacks worldwide; yet the Party laid the foundation by framing COVID as “a nationalist issue—Chinese must stick together and be proud.” Once you invoke the imagined community of the nation, you carve out in-groups and out-groups: the early, real, and positive nationwide solidarity for Wuhan was built on “we are Chinese” rather than “we are all human,” so more radical nationalists hunted for out-groups—Africans / Black people (Guangzhou), foreigners cast as virus carriers, and strong anti-Western sentiment (“the West can’t even handle a pandemic”), which also blinded people to China’s own later missteps (such as the Shanghai lockdowns).
Political Economy: Digital Capitalism and “Enshittification”
The third factor is political economy. He led in with a news story: in 2013–2014, when Kim Jong-un purged his uncle Jang Song-thaek, it was widely reported online that the uncle had been “eaten alive by 120 hunting dogs.” The purge was real (the uncle was most likely executed by firing squad), but the dogs came from a satirical account mocking North Korea; a Hong Kong paper took it as fact, and then the Straits Times, NBC, The Guardian, and USA Today ran it (some with caveats, but with sensational headlines anyway). Why does fake information travel through journalistic circuits? First, North Korea is so inaccessible and seems so “evil” that any absurd detail seems plausible (ironically, its leaders are arguably highly rational—developing a nuclear deterrent against the US is a rational choice). Second, the modern newsroom has broken down: layoffs (in Indonesia, around two-thirds of journalists losing jobs to AI), cost-cutting, reporters no longer on the ground, and outlets copying one another (he admits that in Germany it was common to just translate and copy the BBC). Disinformation campaigns exploit exactly this: plant a story that is often wildly untrue but carries a kernel of truth (the “biolabs” in Ukraine, certain US- or Europe-financed weapon systems); once a paper like The New York Times picks it up, the disinformation actors can say “the Times said so,” lending everything else more plausibility. That is why hollowing out quality journalism is so dangerous—pay your journalists.
He placed all this within “digital capitalism,” an extension of capitalist logic detrimental to how we communicate: the “attention economy” (eyeballs as a commodity to sell ads against), “solutionism” (Silicon Valley claiming more technology will fix technology’s problems), and datafication reaching into every corner of society. He invoked Habermas’s idea of the “colonization of the lifeworld” by the systems of state and market: in everyday interactions—family, friends, students—we treat each other as people, as ends in themselves (the Kantian “never merely as means”); but in these systems we become means to an end—the state’s end is maximizing power, the market’s maximizing profit. Habermas feared these systems would steadily encroach on everyday life; the neoliberalism of Thatcher and Reagan onward pushed public goods—transport, healthcare, education—into market rationality, and now even universities and workplaces are not spared, which is deeply damaging. Capitalism has now expanded into the digital realm through platformization—“the platform society”—dominated by Silicon Valley corporations (Japanese, Korean, and Chinese platforms are growing, but Silicon Valley sets the standard). His interviewees in Taiwan told him: rumors predate the internet, but the internet makes spreading them faster, easier, and more effective.
He bluntly called social media “a massively supercharged weapon for exploiting our cognitive vulnerabilities” (and joked that he personally would like to see Meta and X set on fire). This echoes Cory Doctorow’s “enshittification”: platforms get progressively worse, almost inevitably—because they can easily reallocate value and sit in a two-sided market between buyers and sellers, holding each hostage to the other and skimming an ever-larger share of the value that passes between them. Amazon does this; so do Facebook, X, and Instagram, all to maximize shareholder profit. Musk selling Twitter/X verification badges (letting you “buy your authenticity”) was one failed example. Perhaps we have entered an age in which everything becomes enshittified.
Technology Itself Is Political
The fourth factor is technology. He cited a 2023 case from Taiwan: a rumor spread that the government would issue 100,000 visas to Indian tech workers, instantly igniting extreme xenophobic, racist talk in some circles (“these brown men will come to rape Taiwanese women”). Two issues: first, the original news was false; second, even if it had been true, the level of xenophobic, aggressive racism is the more disconcerting part. And this is bound up with the digital system—Taiwan’s PTT has a downvote mechanic that creates a popularity contest, quite different from caring about what is true or more acceptable, so “the worst rises to the top.” Small-scale environments (LINE, WhatsApp, various groups) are problematic too: people know one another, but if a whole group (and its moderator) is racist, racist discussion persists, unchecked—garbage in, garbage out.
He stressed that technology matters and is itself political. The classic case is Robert Moses’s low parkway overpasses on Long Island in the 1950s: built deliberately too low for buses, to keep poor people, Black people, and Hispanics out of Long Island—racism built straight into the infrastructure (this comes from Langdon Winner’s account; whether the case is fully accurate is debated, but the larger point stands). Our biases—racial, gendered, ableist, classist—get built into infrastructure; technology is in many ways a mirror of society. He also invoked Kranzberg’s dictum: “technology is neither good nor bad; nor is it neutral.” A car can drive your sick grandmother to the doctor or run over protesters—that’s the driver’s fault, not the car’s; but this overlooks that a combustion engine pollutes no matter what you do with it. China wants to replace all combustion cars with electric ones, yet EVs don’t change traffic, waste, or pollution—the better question is why not public transport, why not bicycles? The very idea of the “personal vehicle” carries specific social assumptions. Pro-nuclear narratives are resurging on energy-security grounds (data centers devouring power), but however safe you think nuclear is, it produces waste that won’t disappear for tens or even a hundred thousand years—what are we doing to future generations just so students can “write their essays better”?
He used Bilibili’s “danmu” (bullet comments) as an example: comments scroll across the video, asynchronous yet creating a sense of community and conformity pressure—people say “the next thing expected” in that context, producing a flood of conforming comments. The same is visible across social and digital media: people want to belong, so they post what they think will achieve that, and everyone drifts toward a conformist center; if that conformity is generated by hate or misinformation, hate and misinformation are what get duplicated. He calls this the “viral village”—viral content circulating as if in a village. He noted that McLuhan’s “global village” is often misunderstood: he did not mean “the whole world happily connected,” but that when people get too close they become more savage, impatient, and abrasive—villagers don’t actually love one another (he said this back in 1977).
AI: Don’t Call It “Hallucination”
He then turned to AI—the next hype tech entrepreneurs are selling, packaging generative AI as a wonderful addition to humanity. In the disinformation context, most people worry about two new things: first, deepfakes, fake video and audio convincing enough to pass as real—one good fake clip of “the president” saying “it’ll be fine” in a crisis could cause a costly misjudgment, and the few minutes needed to correct it matter; second, the scaling of “ordinary fakes”—faking is now so easy that, rather than spreading one statement 100,000 times, you can spread 100,000 slightly varied versions that each look like an individual remark, far harder to detect. Fortunately people are also using AI to counter AI (an arms race). But Schneider is more concerned with what AI does to society and the biases within AI—the very backdrop that lets us assess information and that makes deepfakes possible.
He strongly objects to calling AI’s errors “hallucinations”: a hallucination implies a mind that normally tracks reality but is now misreading signals, whereas AI doesn’t care about reality at all—it only computes the probability of words being chained together. He also dislikes “bullshit,” since Trump-style bullshit carries human intent and motivation (even if just narcissistically wanting to look clever), while AI has no mind. He prefers the language of the machine—“AI sewage” or “AI dredge”: stuff flowing out of the machine that settles to the bottom of communication channels and accumulates until the channels are clogged. Treating AI as a “mirror” of society is valuable—it only looks back, only uses the data we feed it, which is exactly why it is racist and sexist: not a bug in the system but the system itself. Others call it “deceitful”—a chatbot is built to feel like a human conversation in order to sell you something; that is the business model, not an accident.
What Can We Do?
His conclusion drew heavily on people he interviewed in Taiwan last year who work on disinformation and digital literacy. The core problem is polarization: people can no longer talk to each other (Taiwan’s green/blue divide, and not only Taiwan). Reconnecting takes years, even decades, of hard face-to-face work: town halls, listening, taking people seriously, and not concluding that anyone who says something “pro-China” must be in the pocket of the communists. A colleague calls this a “whole-of-society approach”—building alliances among civil society, the general public, journalists, politicians, and administrators to chip away at the various issues, including fact-checking; but fact-checking is only one piece of the puzzle, nowhere near enough. We also need to change the script: stop seeing each other as warring opposites, recognize generational differences, and use humor to defuse anger and anxiety (for example, a prime minister stepping in to clarify a rumor that “we are not banning hair products”)—exactly the human, “humor-over-rumor” approach championed by Audrey Tang.
Digital and media literacy matter (we can’t fact-check every single item) and are not mutually exclusive with fact-checking. Ultimately it comes back to treating each other as humans and meeting face to face, reclaiming the colonized lifeworld piece by piece—where transparency, honesty, integrity, and real interpersonal connection are what count. Digital tools aren’t unimportant (one can use AI to counter things too), but they always come with the costs of digital capitalism (intense use of resources, energy, and water) and with the basic limit that “AI is just a mirror and can’t help us do genuinely new things.” On accountability, he argued we must hold responsible not only those who spread content but those who build these environments—above all the tech billionaires. Legal scholars are rethinking whether “crimes against humanity” and “transnational crime” could apply to them (the people running Instagram know its algorithms promote eating disorders among women, yet run them anyway for money). One legal scholar describes the situation as broken technologies and broken governance combining to inflict irreparable harm on the public—and sees it as a possible angle for accountability.
Note taken by Peter Tkach
