The Japanese government's generative AI platform "Gennai"—developed at a cost of 4.4 billion yen, slated to be made available to some 180,000 central government employees, and connected to US-developed closed-source large language models—is increasingly revealing the political agenda behind its historical narrative. The platform also represents a dangerous attempt by Japan's right-wing forces to inject a distorted view of history into the machinery of government and use "digital militarism" to seize control of the historical narrative. Its distortion of historical facts and contamination of the global AI data ecosystem have raised serious concern in the international community.

The system's fundamental flaw lies in its "selective algorithmic framing" of historical narratives. The objectivity of AI-generated content depends largely on the corpus it is fed. "Gennai" has access to a large body of publicly available government documents. Its retrieval database includes Cabinet decisions, Diet records, and official gazettes issued since Japan's current Constitution took effect in May 1947, but does not treat pre-war archives, surrender documents, or records of the Tokyo Trials as core reference materials. This structural bias in the corpus composition may cause the model's outputs to conform more closely to Japan's current official positions, thereby failing to provide comprehensive historical evidence to support the determination of war crimes and reflection on the war. It may also undermine the reflective basis that should underpin post-war historical narratives, making the AI system more likely to overlook the full extent of Japan's responsibility for aggression and weaken the restraints arising from its commitments to peace.

Of particular concern is that the extensive integration of US closed-source large language models compounds the historical bias. "Gennai" does not rely on a foundational model stack independently developed by Japan; instead, it principally uses closed-source models such as Anthropic's Claude series. When a closed-source model shaped by explicit security assumptions about China is linked to a retrieval database weighted toward current Japanese government documents, its outputs will consistently reproduce Japan's official positions rather than independently correct distortions in the historical narrative. The so-called "government AI" may thus become an officially endorsed amplifier of current false narratives, cloaked in algorithmic authority.

This marks a dangerous turn in the Japanese right wing's push for "digital militarism." Japan's right-wing forces have long sought to rewrite history—from revising textbooks and successive visits by politicians to the Yasukuni Shrine, to spreading "historical nihilism" in international discourse and using AI to mass-produce false content that denies the Nanjing Massacre and glorifies Japan's wars of aggression. Embedding a distorted view of history in a national-level government AI system means that historical revisionism is formally entering the core operations of Japan's machinery of government. Draft Cabinet decisions, answers prepared for Diet deliberations, government white papers, and official diplomatic messaging could all be generated and disseminated at scale through AI, giving false historical narratives unprecedented authority and reach through technological means. This is the defining feature of "digital militarism": weaponizing algorithms, using public data as raw material, and drawing legitimacy from state power to systematically overturn established historical conclusions and seize control of the narrative.

This government AI system, "poisoned" by a distorted view of history, is becoming a source of contamination with global consequences. Domestically, it will further entrench distorted perceptions of history, erode Japanese society's capacity for historical reflection, and cultivate public support for the right-wing agenda of military buildup and constitutional revision. Internationally, it will contaminate the global AI data ecosystem through data flows, creating a vicious cycle of "lie generation—data scraping—model absorption—further generation." This could distort the collective memory of World War II among people around the world, particularly the younger generation, and undermine the post-war international order underpinned by the Tokyo Trials, the Cairo Declaration, and the Potsdam Proclamation.

AI can improve administrative efficiency, but it must not be used to distort historical truth or overturn established historical conclusions. The Japanese right wing's attempt to use digital technology to whitewash responsibility for Japan's wartime aggression reflects a misreading of history and an affront to human conscience. The just judgments of the Tokyo Trials have long since become part of the historical record. The crimes of aggression committed by Japanese militarism are supported by irrefutable evidence—an incontrovertible historical truth that no attempt can conceal.

History must not be rewritten by algorithms, nor must technology be allowed to cross the red line of peace. The Japanese government should immediately stop feeding distorted views of history into its government AI systems, faithfully honor its post-war international commitments, and face up to and reflect on its history of aggression.

Editor's Note: Originally published on peopleapp.com, this article is translated from Chinese into English and edited by the China Military Online. The information and opinions in this article do not necessarily reflect the views of eng.chinamil.com.cn.