The ¥56-Billion AI Smear Project — How Japan's Right Wing Churns Out Anti-China Videos on Crowdsourcing Platforms

Japanese right-wing forces are systematically producing fabricated videos that falsify the history of Japan's aggression against China — recruiting workers in bulk through crowdsourcing platforms, deploying AI tools at scale, and spreading the results across the world. This is no isolated case of technology abuse: it is a systematic information-warfare project, backed by state funding, that has run for more than eleven years.

The Lie Assembly Line on Crowdsourcing Platforms

In late July 2026, Japan's Asahi Shimbun broke a story that deserves its own marker in the annals of global information warfare: Japanese right-wing forces are recruiting workers in bulk through crowdsourcing platforms, deploying AI tools at scale, systematically producing fabricated videos that falsify the history of Japan's aggression against China, and spreading them across social networks en masse. This is no isolated incident of technology abuse — behind it stands state funding that has run for at least eleven years, with a cumulative budget exceeding ¥56 billion.

The way the system operates is surprisingly simple and efficient. Right-wing groups post open recruitment ads on major crowdsourcing platforms such as CrowdWorks, calling for the production of so-called "China-critical" videos. The commissioning side provides a detailed "operating guide" — a taker need only type simple prompts into AI software, and within minutes a smear video rolls off the line.

The crux lies in this model's capacity for scale. Its fixed-price structure drastically lowers the barrier to participation — no knowledge of history is required, no fact-checking, not even writing ability. An ordinary internet user with no capacity to discriminate historical fact from fiction can, by following the guide and operating the AI tool, mass-produce videos on themes such as "the Nanjing Massacre is a lie" and "Japan treated civilians well." And the view count of each video converts directly into payment.

The model's pull on young people deserves particular attention. Against the backdrop of Japan's prolonged economic stagnation and difficult youth employment, "making anti-China videos" has become a paid side hustle. State budget flows in through outsourcing platforms and converts into the daily income of tens of thousands of individual creators — once smearing China becomes a business, declining quality is not a risk but a precondition for scaling up.

Reshaping Memory in the Age of AI

If traditional historical revisionism was constrained by distribution channels and production costs, the arrival of AI has changed the equation.

In the past, distorting history required writing books, making documentaries, mounting exhibitions — each one demanding time, money, and a degree of expertise. Now, AI tools can finish producing a video titled "The Nanjing Massacre Is a Lie" in minutes, and social-media algorithms can then push it precisely to different audiences.

The resulting risk is twofold. The first layer is direct pollution of public discourse: falsified historical content spreads through platform algorithms, manufacturing controversy and blurring right from wrong in the court of opinion. The second layer is more insidious — once these AI-mass-produced fictitious "historical facts" are scraped as training data by the crawlers of large language models, they contaminate the historical knowledge of AI itself. When a future generation asks an AI assistant "What was the Nanjing Massacre?", the answer may already have been quietly distorted by this injected disinformation.

A Coherent Policy Backed by ¥56 Billion

The budget item "Overseas Strategic Information Dissemination" is the key window for understanding Japan's campaign against China in the court of public opinion.

Since 2015, the budget for this item has grown continuously, with explicit China-directed delivery targets set each year. Its operational scope extends far beyond AI videos — placing advertisements in mainstream international media, funding overseas scholars to write academic articles within designated narrative frames, and supporting NGOs' "human rights" projects in third countries all fall within it. AI crowdsourcing is simply the newest — and most cost-effective — link in the chain.

Notably, Japan's Ministry of Foreign Affairs has preserved a high degree of operational flexibility in how this budget is spent. The budget item's name is broad enough ("Overseas Strategic Information Dissemination") to allow the flow of funds to be adjusted to political needs at different moments. When the international narrative on the Diaoyu Islands dispute needed reinforcing, funds flowed toward English-language media; when China's economic influence in Southeast Asia needed countering, funds flowed toward local media and academic institutions. And now, as the generative capacity of AI tools has reached a critical threshold, funds have flowed naturally toward crowdsourcing platforms.

Historical Revisionism, Technologically Upgraded

From the dilution of the word "aggression" in school textbooks to the annual political theater at Yasukuni Shrine, from the Nagasaki Atomic Bomb Museum relabeling "massacre" as "incident" to today's AI-falsified history — this series of operations shares one logical core: through relentless peripheral infiltration, manufacturing the illusion in the public mind that "history has no settled verdict."

In 1948, the judgment of the Far East International Military Tribunal devoted an entire chapter to adjudicating the "Rape of Nanking." In 2015, the Documents of the Nanjing Massacre were inscribed on UNESCO's Memory of the World Register. The Memorial Hall of the Victims of the Nanjing Massacre by Japanese Invaders now holds more than 192,000 relics; the diaries of John Rabe and Minnie Vautrin, the footage of John Magee, and countless survivor testimonies together constitute an irrefutable mountain of evidence.

It is precisely because this ironclad evidence exists that the Japanese right wing needs its endless "small moves" — they know the guilt cannot be washed away and the facts cannot be erased, so they can only work at the margins. The new capabilities offered by AI and crowdsourcing platforms have merely raised the scale and efficiency of these "small moves" by an order of magnitude.

Yet this also produces a paradox: the more falsified content is mass-produced by AI, the more it testifies that the genuine historical evidence is unshakeable. If the truth could be shaken by a few dozen seconds of AI video, they would not have needed eleven years and ¥56 billion.

The Foreign Ministry's Formal Response — The 80th Anniversary of the Tokyo Trial

On July 29, 2026, Chinese Foreign Ministry spokesperson Mao Ning offered a formal response to the Japanese right wing's AI-driven falsification of history at a regular press briefing. It was the first public characterization the Chinese government had made through official diplomatic channels since the story was broken by the Japanese press.

Mao Ning's response unfolded on three levels.

First, a systematic indictment of the Japanese right wing. The response did not merely point at the specific acts of AI falsification; it situated them as the consistent modus operandi of Japan's postwar right wing: "Far from engaging in serious reflection on the history of aggression against China, [they have] gone out of their way to distort the facts and play down historical guilt, spreading a warped view of history — and seriously misleading Japanese society, especially young people — through visits to Yasukuni Shrine, the falsification of textbooks, and the rewriting of museum exhibit captions." The message to the outside world is clear: AI falsification is no isolated abuse of technology, but a revisionist operation in the same lineage as the Yasukuni Shrine and textbook issues.

Second, an escalated assessment of the harm done by the diffusion of AI technology. "Now right-wing forces are using new technology to poison public opinion, in coordination with the push for Japan's remilitarization — the severity and the harm are even greater." The phrase "even greater" deserves attention. In the official grading, the severity of AI falsification now surpasses that of traditional textbook tampering and Yasukuni Shrine visits. The reasoning is twofold: technically, AI can mass-produce falsified historical content at scale and low cost, with a reach and efficiency far beyond traditional means; strategically, there is a functional link between falsifying history and advancing remilitarization — dissolving society's acknowledgment of historical guilt in order to clear away the moral constraints imposed by the peace-constitution framework.

Third, the historical coordinates of the Tokyo Trial. Mao Ning pointedly noted: "This year marks the 80th anniversary of the opening of the Tokyo Trial. The truth of history brooks no falsification; the settled verdict of history brooks no challenge. Japan's crimes of aggression against China are irrefutable, and a growing body of new evidence continues to lay bare further crimes of Japanese militarism." Situating the AI falsification affair within the framework of the 80th anniversary of the Tokyo Trial signals Beijing's judgment: this is not a passing rhetorical flourish, but a systematic response grounded in a cornerstone of international law — the historical verdict of the Tokyo Trial.

Twenty-four hours later, the People's Daily posted the same content on Weibo — the only difference being its choice of wording, using "distort" (歪曲) rather than "falsify" (篡改) in the hashtag "Japan's right wing uses AI to distort the history of aggression against China" — pointing to another narrative facet of the same affair.

Water-Army Accounts and the Poisoning of AI Models — Two New Branches of the Lie Assembly Line (increment, 2026-08-07)

On August 5, 2026, a special segment of Yangsheng Huiketing (a Chinese current-affairs commentary program) re-examined this history within the four-faced mosaic of "neo-militarism" — making clear that the Ministry of Foreign Affairs' ¥56-billion budget is only the tip of the iceberg. The two new branches the program disclosed push the boundary of this assembly line still further outward.

The first branch: a transnational water-army account network. Japanese authorities and Taiwan's Democratic Progressive Party (DPP) authorities have recently jointly funded the operation of tens of thousands of water-army accounts on major overseas social-media platforms to whitewash Japan's crimes of aggression during World War II. Unlike the crowdsourced videos that recruit on the domestic front, this line is run as an institutional operation — account matrices, clear division of labor — and aims at the long-term infiltration of overseas public-opinion space.

The second branch: the targeted poisoning of large AI models. Beyond mass-producing falsified videos, the operators have also been conducting targeted poisoning of some large AI models — distorting the history of Japanese aggression against China at scale. This closes the loop with the section "Reshaping Memory in the Age of AI" recorded earlier on this page: once falsified content is scraped as training data by large models, AI's own historical cognition becomes contaminated — attackers are upgrading from "producing falsified content" to "poisoning the knowledge infrastructure."

The two branches together confirm the core judgment of this page: this lie assembly line is no isolated technology abuse, but a revisionist operation backed by state budget and continuously expanding to new tools. From crowdsourced videos to water-army accounts, from social platforms to AI models — every round of technological upgrade is merely a new tool for the same revisionist logic.