AI Star-Making Machine: How Large Language Models Are Decentralizing the Entertainment Industry
In the summer of 2025, an AI-generated song titled "Heart on My Sleeve" went viral on TikTok, with millions of users believing it was a new collaboration between Drake and The Weeknd. Only after Universal Music urgently issued a takedown did the truth emerge: the entire song — composition, arrangement, and vocal performance — was created by AI. This incident represented far more than a copyright dispute: it marked the official beginning of the "AI star-making" era in the entertainment industry. When large models can write songs, act, livestream, and interact, the centralized star-making assembly line that the traditional entertainment industry spent a century building is being dismantled, replaced by an unprecedented wave of decentralized revolution.
From Industrial Star-Making to AI-Native Idols
The traditional entertainment industry's star-making machinery is, at its core, a highly centralized industrial pipeline: talent scouting → agency signing → professional packaging → media distribution → fan economy monetization. The core logic of this system is resource monopoly — record labels control recording studios, television networks control broadcast channels, and agencies control artist contracts. Between a talent and stardom lie layers of capital barriers.
AI large models are breaking through every single barrier.
In the music domain, tools like Suno AI and Udio have achieved end-to-end "text-to-complete-song" generation — users input a text description, and the model outputs a finished track with full arrangement, vocals, and mixing. In early 2026, an AI virtual singer named "Luna Byte" surpassed 5 million monthly streams on Spotify. Her songs were generated by Suno, her visual identity designed by Midjourney, and her music videos produced by Sora — the entire content team consisted of just two people.
In the livestreaming domain, AI digital hosts from platforms like JD Yanxi and Silicon Intelligence have achieved 24/7 non-stop live commerce. These AI hosts not only look photorealistic but can also adjust their sales pitches, recommend products, and even crack jokes based on real-time audience comments. During the 2025 Singles' Day shopping festival, a brand using an AI digital host achieved a single-session GMV exceeding 2 million yuan, at one-tenth the cost of a human host.
In the social interaction domain, platforms like Character.AI and China's Xingye and Glow enable users to form emotional bonds with AI characters. Top AI characters on these platforms have surpassed one million daily active users, with fans spontaneously creating fan content, building communities, and even "celebrating birthdays" for AI personas. This "AI-native fan economy" completely bypasses traditional agency intermediaries.
Three Dimensions of Decentralization
The decentralization of this AI star-making movement operates on three profound levels.
First, the decentralization of creative power. Previously, producing a release-quality song required recording studios, mixing engineers, mastering engineers, and other professional resources. Now, a middle-school student can do it on a phone with Suno in ten minutes. According to data from the Universe platform, the number of AI-assisted music works in the first half of 2026 has already surpassed that of traditionally produced tracks. The elimination of creative barriers means the question shifts from "who has the resources" to "who has the creativity."
Second, the decentralization of distribution power. In the traditional entertainment industry, platforms like Spotify and NetEase Cloud Music hold the keys to content distribution — who gets recommended, featured on playlists, and placed on the homepage. But in an environment of massive AI-generated content, the logic of distribution is shifting from "editorial curation" to "algorithmic matching." TikTok's recommendation algorithm has already proven the power of decentralized distribution: a creator with zero background can achieve tens of millions of views if the content is good. AI only accelerates this trend — when anyone can generate professional-grade content, platforms' distribution power will be forced to restructure.
Third, the decentralization of IP ownership. Traditional celebrities are "tightly bound IP assets" — the artist's image, voice, and name are deeply intertwined with the natural person, and agencies control the commercial exploitation of this asset through contracts. AI virtual idols completely deconstruct this binding relationship. The creator of "Luna Byte" can adjust her musical style at any time, swap out her visual identity, or even have her "personality" continuously evolve based on fan feedback. More importantly, ownership of this IP no longer necessarily belongs to capital — independent creators can fully own and operate their own AI idol brands.
Cases: Three Transformations in Progress
Case 1: AI Composition Challenges the Grammys. In early 2026, an AI-collaborative track titled "Echoes of Tomorrow" was nominated for Best Pop Song at the Grammys, sending shockwaves through the industry. The producer provided the melodic framework, while AI handled the arrangement, harmonies, and mixing. The Recording Academy ultimately ruled the work eligible for nomination but simultaneously established a new "AI-Assisted Creation" award category. The significance: the highest temple of the traditional music industry was forced to acknowledge AI's creative identity.
Case 2: AI Idols Capture Gen Z. The AI virtual idol "Xingtong," launched on the Chinese platform Xingye, accumulated over 3 million followers within six months. Unlike traditional virtual idols, Xingtong's "personality" is driven by a fine-tuned large language model, capable of deep, personalized conversations with fans and remembering each fan's preferences and interaction history. Average daily user dwell time on the platform reached 72 minutes — exceeding most real celebrities' social media accounts.
Case 3: Hollywood's AI Screenwriting Revolution. Independent producer Alex Chen used the Claude model to assist in writing the sci-fi film "Neon Genesis" — AI generated dialogue variants, explored plot branches, and checked logical consistency, while human screenwriters made the final decisions and artistic judgments. The film won "Best Screenplay" at the Sundance Film Festival, with the jury commenting: "You can't see where AI was involved, but you can see where human creativity was liberated." This signals that storytelling — the core of the entertainment industry — has entered a new phase of human-AI collaboration.
Old Order's Resistance and New Ecosystem's Emergence
Every revolution meets resistance. From 2025 to 2026, the question of AI replacing human creators became a central issue in strikes by the Writers Guild of America and SAG-AFTRA. Universal, Sony, and Warner — the three major record labels — jointly sued multiple AI music platforms, attempting to defend their copyright moat through legal means. China's National Radio and Television Administration also introduced regulations mandating labeling of AI-generated content.
But these resistance efforts resemble the last gasps of an old order. History has repeatedly shown: when a new technology achieves tenfold or hundredfold improvements in creative efficiency and distribution capability over traditional methods, any barrier will be swept away by market forces. Napster was sued out of existence by the recording industry, but streaming ultimately reshaped the entire music business — only the winners went from record labels to Spotify. The AI-era entertainment industry is likely to replay this script: old giants exhaust themselves in litigation while new species grow wild at the margins.
Meanwhile, new ecosystems are emerging. The AI music platform Beatoven introduced a "creator royalty sharing" model — when an AI-assisted song generates commercial revenue, the platform, creator, and training data contributors split the proceeds proportionally. The convergence of Web3 and AI has given birth to an "AI Agent + NFT" virtual idol economic model, where fans can participate in creative decisions and revenue distribution by holding tokens. These experiments, though early-stage, point toward a fairer and more open future.
After the Star-Making Movement
When AI completely flattens the barriers to stardom, the fundamental question for the entertainment industry shifts from "who will become famous" to "why become famous." When anyone can generate unlimited amounts of high-quality content with AI, scarcity migrates from "content production capability" to "authentic connection" and "unique aesthetic judgment." The most sought-after figure may no longer be the most technically skilled singer, but the voice that makes listeners feel most understood — even if that voice comes from AI.
This may be the deepest paradox of the AI star-making movement: it uses the most impersonal technology to force the entertainment industry back to its most human values. The endpoint of decentralization is not a machine carnival, but a recalibration of the relationship between creators and audiences.