By Lu Hao
Recent media reports have revealed that Japanese right-wing forces have been feeding falsified material about World War II into leading international large language models (LLMs), causing some models to produce responses that reflect clearly distorted views of history. As these forces move from falsifying textbooks and visiting Yasukuni Shrine to systematically "poisoning" artificial intelligence (AI) training data, a battle for minds is already underway. The danger lies not only in the lies themselves, but also in the way they are packaged as "algorithmic conclusions," allowing them to insidiously permeate society and steadily erode humanity's collective historical memory.
"Data poisoning" by Japanese right-wing forces has long ceased to consist of isolated acts. It has developed into a highly organized, well-funded operation supported by technology. Multiple sources indicate that Japanese right-wing forces are openly recruiting video creators to smear China, offering substantial payments for fabricated videos. Since 2015, Japan's Ministry of Foreign Affairs has poured substantial funding into this illicit operation under the banner of strategic overseas communications. By setting explicit rates and paying for each item produced, it has helped turn the fabrication of historical falsehoods into a mass-production business.
More alarming still, this practice is moving beyond privately run illicit operations to become a state activity. "Gennai," a government AI system developed under the leadership of Japan's Digital Agency, was made available to government personnel in 2026. The system sets May 1947, when Japan's postwar pacifist Constitution came into force, as the starting point for collecting its source material, excluding from its core corpus material on the Sino-Japanese War of 1894–1895, Japan's war of aggression against China, colonial rule, and even the Tokyo Trials judgment. This means that when Japanese government personnel use Gennai to draft policy documents, the system automatically produces a carefully edited historical narrative. Meanwhile, this AI-generated content, bearing an official stamp, is collected for inclusion in datasets worldwide as "highly credible" source material, creating successive layers of data contamination.
Japanese right-wing forces are extending their reach into the rapidly developing field of AI, seeking to slip carefully packaged distortions of history into vast amounts of training data. They then exploit LLMs' "black box" nature and tendency to "blindly accept" such material, allowing falsehoods to insidiously distort public understanding. The danger must not be underestimated. These tactics are covert: ordinary users struggle to detect that the information has been contaminated. They are persistent, enabling long-term infiltration through systematic, engineered processes. And they are deceitful, falsifying history in the name of technological progress. From tampering with written records to "poisoning" online algorithms, the sinister intentions of Japanese right-wing forces remain unchanged.
"Data poisoning" is far more than a misuse of technology; it is a carefully orchestrated campaign cognitive warfare. Its true purpose is to distort global understanding of World War II history and the crimes of aggression committed by Japanese militarism, paving the way for Japan to cast off the constraints of the postwar order and accelerate its "remilitarization." This campaign strikes at the very foundations of historical justice and challenges the fundamental moral boundaries of humanity's conscience.
In recent years, the Japanese right-wing forces have pushed relentlessly to revise the pacifist Constitution, move beyond the "exclusively defense-oriented policy," and openly expand the armed forces and their arsenals. All these moves depend on downplaying or even whitewashing Japan's responsibility for its past crimes. They seek to make the world forget the facts of Japanese aggression and enable Japan to become, in their words, "strong again" politically and militarily, thereby making room for the specter of militarism to return.
By turning AI into an accomplice in distorting the truth and whitewashing their country's culpability, Japanese right-wing forces are not only deeply hurting the feelings of people in countries subjected to Japanese aggression, but also shamelessly trampling on historical truth and legal justice. In the face of these new tactics for falsifying the history of Japanese aggression, the international community must stay vigilant work together to resist them, strengthening collective defenses against cognitive manipulation in the digital era.
(The author is director of the Comprehensive Strategy Research Office at the Institute of Japanese Studies, Chinese Academy of Social Sciences)