Shen Anyu's cloned voice narrates content he never recorded, and platforms now flag his real work as AI-generated — suppressing his views and income. He has filmed himself proving he is human repeatedly and taken the case to court.
2026-07-11 · by Moe Ameen
Shen Anyu, a 31-year-old professional voice narrator from Xuzhou in Jiangsu province, has spent months fighting AI copies of his own voice — and, in a twist, proving to platforms that his real recordings are actually his. As reported by Sixth Tone, unauthorized AI clones of Shen's voice have spread widely since 2025, narrating movie explainers, sports news, product promotions, conspiracy videos, and short clips he never recorded. As the synthetic versions multiplied, platform detection systems began flagging his authentic recordings as AI-generated, which suppressed his reach and cut his income.
To keep working, Shen has repeatedly recorded verification videos to demonstrate he is a real person, in one case reciting a tongue twister to show his genuine voice. His wife, Wei Yiyuan, has spent significant time documenting the fakes, filing platform complaints, and gathering evidence — an effort she compared to "punching into cotton." Shen built his career narrating for a Douyin channel with millions of followers, at his peak earning a solid monthly income from the work, and watched bookings decline as collaborators turned to cheaper cloned voices.
Shen took the dispute to court. A first hearing was set for July 2, 2026, but the defendant moved to transfer the case to a court in another city before it could begin. To prevail, Shen would need to establish a clear evidentiary chain showing the synthetic voice was derived from his own recordings, which could require forensic voice analysis costing at least 10,000 yuan. His lawyer framed the case as part of China's still-forming jurisprudence on AI voice infringement. The problem is industry-wide: other voice workers report rates halving and suspected unauthorized cloning, in a market serving hundreds of millions of audio-content users.
Shen's ordeal is a warning about where a voice comes from, and it points to the honest way to use synthetic voice at all. The harm here is unconsented cloning of a real person: someone scraped his recordings, generated a convincing double, and flooded platforms until his own work got mislabeled. That is the opposite of how a content engine should operate. Kompozy is a content generation and publishing engine, and the video it produces — Persona Shorts, Persona HeyGen, Persona Frames, Marketing Shorts — runs on HeyGen's avatar and native voice inside an AI Influencer persona you set up and control, not a lifted clone of a stranger. You build one consistent identity, governed by a Persona Brief, and reuse it across everything. The voice and face are an asset you own the creation trail for, which is the exact provenance Shen has to spend 10,000 yuan in forensics to prove after the fact.
There is also a same-week angle for creators covering this. "A voice actor is being forced to prove he's human as AI clones flood his feeds" is a story your audience is reading right now, and it maps to a real editorial position: consent, disclosure, and owning your identity. Drop your take into Kompozy and it fans one point of view into a Blog Article explainer, a captioned Persona Short, a brand-exact Carousel, a Quote Graphic pulling the "it erases the person behind the work" thread, and platform-native Text Posts — all held to your voice by the Persona Brief. Autopilot then schedules and publishes the set across nine social platforms plus blog and email from one queue, so you ship a clear, human position on AI voice cloning while the news is fresh.
Chinese voice narrator Shen Anyu found that AI clones of his voice had spread so widely across platforms that detection systems started flagging his own authentic recordings as AI-generated. Because a mistaken "AI" label suppresses reach and income, he repeatedly filmed himself to prove the recordings are genuinely his, in one case reciting a tongue twister.
He took the case to court. A first hearing was scheduled for July 2, 2026, but the defendant moved to transfer it to a court in another city before it began. To win, Shen would need to establish that the synthetic voice was derived from his recordings, which could require forensic voice analysis costing at least 10,000 yuan. His wife has separately documented fakes and filed platform complaints.
The harm here is unconsented cloning of a real person's voice. Legitimate synthetic voice comes from a persona you set up and control, with a clear creation trail. A content engine like Kompozy generates video on a HeyGen-driven AI Influencer persona you own rather than a scraped clone of someone else, so the identity is defensible and consistent by design.
That being human is not automatically legible to a platform's AI detector, and that proving provenance after the fact is slow and costly. The practical response is to own a consistent, consented identity and keep a clear creation trail, and to disclose AI use rather than pass synthetic media off as an uninvolved real person.