Japan's lower house election becomes a testing ground for generative AI misinformation

Japan's voters are grappling with a flood of AI-generated fake content during the lower house election.

Fake videos and fabricated news are spreading rapidly across social media, Japan Times reports. Generative AI now makes it easy to create convincing fakes, and platforms like YouTube and TikTok reward high engagement, creating a perfect storm for misinformation.

One fake video showed two party leaders alongside a communist-style red logo. A YouTube channel featuring AI-generated grandmothers ranting about politics racked up nearly six million views. According to Professor Shinichi Yamaguchi from the International University of Japan, 51.5 percent of respondents in a comprehensive survey believed fake news to be true—a threat to democracy.

The problem cuts both ways. A candidate posted video footage of a large crowd at his election rally. The AI chatbot Grok flagged it as AI-generated, even though it was authentic.

This phenomenon is known as the liar's dividend: the mere existence of generative AI gives liars a convenient excuse to dismiss real evidence as fake. Donald Trump used this tactic when he claimed that authentic photos of a Harris rally were AI-manipulated ("AI'd"). Beyond Japan, primarily conservative groups are deploying AI in election campaigns—sometimes openly, sometimes not.