The Asahi Shimbun, one of Japan’s largest newspapers, reported on how US campaigns are working to shape what AI chatbots tell voters about candidates. The article quotes Civly CTO Matthew Hanauer.
Original articles
- Online, October 12, 2026: 選挙で勝つには有権者よりAI説得、でも利用は隠しがち 米中間選挙 (“To win an election, persuade the AI more than the voters, though campaigns tend to hide it: US midterms”). By Takeshi Narabe, San Francisco Bureau Chief. Part 6 of the series “From the Ground: US Midterms 2026.”
- Print, October 15, 2026, morning edition, page 3: （アメリカ中間選挙２０２６）ＡＩを味方につけろ 政策や人物、思い通りの回答が出るよう「最適化」 (“US Midterms 2026: Get AI on your side, ‘optimizing’ so policies and people come out the way you want”). A shorter version of the online article.
Both articles are in Japanese. The full text is available to Asahi Shimbun subscribers. The articles write Civly as シブリーAI and Mr. Hanauer’s name as ハノウアー.
English translation of the opening
Whom should I vote for? More and more voters are asking AI (artificial intelligence) to help make that important decision. How AI presents information about candidates could sway the results of elections. The November 3 US midterm elections are a campaign in which candidates and voters are using AI more widely than ever before.
According to the Pew Research Center, the share of adults who use AI chatbots rose from about 33% in 2024 to about 49% in 2026. How can candidates and companies get AI to give the answers they want to put forward? As more people use AI, demand is growing for businesses that give that kind of advice. It is called “AI answer optimization” (AEO).
“Voters use AI to look up candidates, so candidates want AI to show them accurately and as they intend,” says Matthew Hanauer, chief technology officer of US-based Civly AI, which runs an AEO business.
The free preview ends here. The rest of the article, about 1,900 Japanese characters, is available on asahi.com.