Esquire Singapore published what appeared to be an interview with Mackenyu Maeda, the actor who plays Roronoa Zoro in Netflix's live-action One Piece. It wasn't. The publication used Claude and Microsoft Copilot to generate the entire thing, feeding in text from Mackenyu's previous interviews and letting the AI produce new responses.
The credited writer, Joy Ling, included a disclosure that reads more like a boast: "We were stoked to have some face time with the Japanese-American actor, but his schedule prevented it. With a driving need for a feature, we had to be inventive." Inventive is one word for it. The AI-generated answers included the actor supposedly discussing his deceased father, legendary action star Sonny Chiba, and expressing a desire to make him proud. An AI hallucinated emotional quotes about a real person's dead parent, and an editorial team approved it for publication.
The article was published in early March 2026 but only drew widespread attention in April when fans flagged it. One fanpage response summed up the reaction: "I'm disappointed Esquire SG wrote an entire AI interview to replace Macken's response."
This is a clean case study in where AI content generation crosses a line. Using AI to summarize research, draft outlines, or assist with copy is one thing. Manufacturing quotes and attributing emotional statements to a real person is fabrication, regardless of the tool used to do it. The "we edited the AI output afterward" defense doesn't hold up when the core problem is that the subject never said any of it.
For anyone building content workflows around AI tools, the lesson is straightforward: AI can help you write faster, but it cannot conduct an interview. The moment you present generated text as someone's actual words, you've left journalism and entered fiction.