AI Tools: The ElevenLabs Situation Is Crazy

Watch on YouTube: https://www.youtube.com/watch?v=lxDNHLgWInU

Channel: Chris Williamson

Introduction

In a recent discussion, Chris Williamson described how his own voice appeared on the Eleven Labs platform as a synthetic clone named “Archer” without his permission. The incident raises broader questions about voice ownership, the reliability of technical similarity tests, and the legal landscape that currently protects—or fails to protect—individuals from unauthorized AI‑generated voice use. The following breakdown extracts the key points from the video, presents the technical details shared, examines the limitations of the current system, and outlines practical steps anyone can take if they encounter a similar situation.

What the video covers

Williamson begins by recounting the moment he discovered the Archer voice on Eleven Labs, noting that the synthetic voice not only matched his timbre but also reproduced his idiosyncratic speech patterns. He explains that Eleven Labs responded with a similarity score from an ENCAPA‑TDNN test, which fell below their internal threshold for a confident match. Despite the technical result, Williamson points out that listeners perceived the clone as highly similar, suggesting a potential right‑of‑publicity concern. The video then moves to the legal context, citing precedents such as the Bette Midler and Tom Waits cases, while emphasizing that no federal statute protects a person’s voice; protection relies on a patchwork of state right‑of‑publicity laws. Williamson shares advice from a copyright lawyer on gathering evidence and pursuing a claim, notes the subsequent removal of the Archer voice (though the reason remains unclear), and warns that the current redress mechanism—primarily a support email and unverifiable tests—leaves individuals vulnerable. He concludes with a call to treat voice as intellectual property, monitor for unauthorized clones, and advocate for clearer legal safeguards.

Step‑by‑step walkthrough of the Eleven Labs situation

• Discovery: Williamson finds a voice labeled “Archer” on Eleven Labs that sounds strikingly like his own, including personal speech quirks.
• Initial response: Eleven Labs provides an ENCAPA‑TDNN similarity score of 0.28, which is below their 0.7 threshold for a high‑confidence match.
• Perception check: Despite the low score, Williamson and listeners report strong auditory similarity, raising concerns about how the voice is perceived by the public.
• Legal consultation: He speaks with a copyright lawyer who advises documenting any commercial or misleading use, collecting screenshots, timestamps, and listener comments that note the resemblance.
• Evidence gathering: Williamson outlines the types of proof needed—commercial usage examples, listener testimonials, and any metadata that could trace the voice back to the platform.
• Platform action: The Archer voice is later removed from Eleven Labs, although it is uncertain whether the removal resulted from his complaint or other factors.
• Reflection: Williamson discusses the broader implications, noting that voice is as personal as a fingerprint and that unauthorized synthetic copies can erode trust and enable false endorsements.
• Advocacy: He urges viewers to treat voice as intellectual property, monitor for clones, and push for clearer legislation governing synthetic voice use.

Tools, numbers and technical details

• Platform: Eleven Labs (AI voice‑generation service).
• Similarity metric: ENCAPA‑TDNN test.
• Reported score for the Archer voice: 0.28.
• Eleven Labs internal threshold for a high‑confidence match: 0.7.
• Legal precedent reference: Tom Waits case resulted in a $2.4 million settlement.
• Illustrative cost analogy: Williamson compares the ease of cloning a voice to the price of a cheeseburger, emphasizing low barrier to entry.
• Other cited figure: $200 million spent on MH370 searches (provided as contextual background, not directly related to the voice issue).
• Example of a misleading product price mentioned in the video: $10,000 for a fake “Loco” product.

Caveats and hype check

While the video highlights how inexpensive and accessible voice cloning has become, it does not present systematic data on the scale of misuse or revenue generated from unauthorized clones. The claim that voice cloning can be done “for the price of a cheeseburger” serves as an illustrative metaphor rather than a verified market price. Williamson’s concerns about widespread abuse remain anecdotal; the absence of concrete statistics means the extent of the problem is not quantified in the presentation. Additionally, the similarity test used by Eleven Labs is presented as a proprietary metric, and the video does not disclose the methodology behind the 0.7 threshold, leaving room for uncertainty about how definitive the test truly is.

“Your voice is the vocal chords equivalent of a fingerprint.”

“If you see or hear the same person’s AI clone enough times, you’ll start distrusting the actual person even when they haven’t done anything wrong.”

“There is currently no referee on Earth who can settle whether a voice is yours or not.”

Concrete next actions

• Document any instance where your voice appears to be used commercially, in advertising, or in misleading contexts; capture screenshots, timestamps, and URLs.
• Collect listener feedback or comments that note the similarity between your natural voice and the suspected AI clone.
• Consult a lawyer experienced in right‑of‑publicity, intellectual property, or entertainment law to evaluate the strength of a potential claim.
• If litigation becomes a possibility, request formal discovery from the AI platform for similarity test data, training source documentation, and any internal voice‑cloning logs.
• Regularly monitor voice‑cloning marketplaces and catalogs for unauthorized clones of your voice; issue takedown notices under the platform’s policy or applicable law when appropriate.
• Consider licensing your voice through official channels or reputable voice‑licensing services to retain control and create a potential revenue stream.
• Advocate for clearer legal protections by contacting elected officials, supporting bills that address synthetic voice rights, or participating in industry discussions about ethical AI use.