虎嗅

AI Drafting Using the 101-Copy System: Chaos Everywhere, Yet No One Has Debuted Yet

原文:复制101赛制的AI选秀:喧嚣遍地,无一人出道

Summary of Key Points

After the pause in traditional talent shows, AI-based talent competitions have been seen by investors as a perfect alternative due to their perceived advantages—no scandalous incidents, lower costs, and fewer disputes. However, current attempts (such as TikTok-style mass-produced variety shows and Bilibili creator-hosted events) have not been successful. TikTok shows rely on catchy tags (like zodiac signs and animal years) to attract attention, but they essentially involve role-playing, failing to retain fans who are interested in the contestants. Bilibili events target the anime and manga community, and while they have produced a few AI idols with followings, it is difficult to replicate this model on a large scale. AI talent shows generally lack the core appeal of traditional shows, including the sense of competition, uncertainty, and engagement from fans, as well as the need for well-designed concepts and sustained operations. There is still a long way to go before we see truly successful AI idols.

I. Two Approaches to AI Talent Shows, Both with Challenges

There are two main approaches to AI talent shows, but both have significant flaws:

  • TikTok-style Mass-produced Variety Shows: Playing the “Pseudo-Talent Show” Game

Examples include “12 Star练” (which personifies zodiac signs) and “Starlight 108” (108 AI girls). These shows mimic the format of traditional talent shows, but the essence is different:

  • Fans of “12 Star练” vote not because they like the contestants, but to support their zodiac sign; after watching, they leave without a sense of fulfillment.
  • “Starlight 108” is more like an interactive short drama, where the contestants’ fates are predetermined, and viewer votes cannot change the plot, leading to feelings of being deceived.
  • These shows are highly homogeneous (with themes like the 12 zodiac signs and MBTI personality types), and their popularity fades quickly (the likes for “12 Star练” dropped from 700,000 to 50,000-70,000 later on).
  • Bilibili Creator-hosted Events: Targeting the Anime and Manga Community, but with Limited Success

Examples include “AI Idol Debut Season,” which is more of a creator recruitment event for the anime and manga audience. While it produced a somewhat successful AI idol like EIKO, with a music video that resonated with young people about workplace stress and AI-related anxieties, the issues are clear:

  • The popularity is concentrated among a small group of fans.
  • After the events, the contestants rely on the creators for further promotion, lacking sustained support and the ability to create a large number of stars.

II. Why Don’t Fans Buy In? AI Talent Shows Lack the “Soul” of Traditional Shows

Investors think fans only care about attractive faces and no scandals, but the appeal of traditional talent shows goes beyond that:

  • Lack of Competition: In traditional shows, there is a complex interplay between companies and fans, who have to discern between “scripted” and “elected” contestants and compete for resources. This lack of interaction is evident in AI shows (e.g., voting on TikTok is just liking, without the underlying social dynamics).
  • Lack of Uncertainty: The real-life contestants’ reactions (crying under pressure, adapting after mistakes) are genuine, whereas AI contestants’ fates are predetermined.
  • Lack of Fan-created Relationships: Fans enjoy creating their own relationships and fan fics, but AI contestants’ relationships are pre-set, leaving no room for such interaction.

These elements are essential and cannot be replicated by AI shows, leading to fans losing interest quickly.

III. The Secret to Creating Successful AI Idols: High Concepts and Engaging Narratives

While AI can handle image generation, creating memorable idols requires a strong narrative and aesthetic:

  • Example 1: EIKO: Her music video has a consistent cyber aesthetic and focuses on the anxieties of the AI era, providing a compelling story that helped her gain a small following.
  • Example 2: K-pop Group aespa: Although they are a real-life group, their success came from a high-concept video featuring both real and AI characters exploring the wilderness, with a cinematic narrative and continuously updated storyline.
  • Domestic Shortcomings: Most AI idols lack such a strong narrative and aesthetic; they either rely on clichéd themes (zodiac signs) or tired plots, and lack the support of dedicated teams to refine their content.

IV. The Illusion of Profitability for Investors

Investors initially thought AI talent shows would be highly profitable (no need for profit-sharing, mass production), but reality is different:

  • Difficulty in Maintaining Popularity: TikTok shows see initial spikes followed by declines in viewership.
  • Challenges in Monetization: Even successful shows like “12 Star练” struggle to attract long-term brand partnerships due to lack of uniqueness.
  • Talent Gap: AI cannot solve the issue of creating compelling narratives; there is a shortage of top-tier creative teams (like SM Entertainment, which produced aespa) and regular performance opportunities.

Only when investors focus on content and operations rather than seeing AI as a shortcut can AI idols become more than temporary fads.

Conclusion

AI talent shows are currently just a bubble. To succeed, they need to address the core needs of fans by reinvigorating the sense of competition and engagement or by creating engaging narratives for the anime and manga community. This requires long-term investment, not short-term traffic generation. The illusion of AI being a perfect substitute for traditional talent shows must be grounded in high-quality content.