Bill Kurtis Found 200-Plus AI Copies of His Voice — and He Is Suing
Broadcaster Bill Kurtis found over 200 unauthorized AI clones of his voice, fueling 13 class actions against Google, Meta, Nvidia, Apple and ElevenLabs.
Why it matters
- Bill Kurtis found more than 200 AI-replicated voice-overs using his voice without consent, after a viewer email on Oct. 1.
- Attorney Ross Kimbarovsky of Loevy & Loevy has filed 13 proposed class actions against Google, Amazon, Meta, Nvidia, Microsoft, Apple and ElevenLabs.
- The lawsuits rest on two Illinois laws covering AI 'digital replicas' and AI voice-model training without notice or permission.
- The No Fakes Act, backed by SAG-AFTRA, cleared the U.S. Senate Judiciary Committee and awaits a full Senate vote.
- Kurtis counts almost 1,000 narrations across a 60-year broadcasting career.
Veteran television journalist Bill Kurtis says he has found more than 200 examples of his own voice replicated by artificial intelligence — none of them made with his knowledge, and none with his consent. The figure comes from his own reporting, triggered by a single viewer email dated Oct. 1, and it now sits at the center of an escalating legal fight over AI voice cloning that has drawn in Google, Amazon, Meta, Nvidia, Microsoft, Apple and ElevenLabs.
Kurtis, a former CBS network and WBBM-Ch. 2 news anchor who has spent 60 years in broadcasting, laid out his case in a first-person article published by the Chicago Tribune and distributed by Project Syndicate. His count of unauthorized voice-overs is the story's hardest number. His framing is blunt: the technology threatens his livelihood, his career and his journalistic reputation, and he refuses to surrender any of them.
How one viewer email uncovered 200 cloned voice-overs
Kurtis describes himself as a narrator — someone who, in his words, transforms "the written word into audio that matches the visuals." Across a 60-year career, he counts almost 1,000 narrations spanning news, true crime programs, nature documentaries and special events.
The discovery began on Oct. 1, when a viewer emailed him about a cable show voiceover she suspected was synthetic. Her message, which Kurtis quotes verbatim, reads: "the voiceover, I believe (is) artificial intelligence (AI), and it sounds, I mean, I don't know if Bill Kurtis said that, great it's awesome, I love him. And yet, I don't know that that's true, so I thought it was AI and I would inform him, because I don't think that's okay, replicate someone's voice without their permission."
Kurtis calls that single sentence a capsule portrait of "a crisis in the news and entertainment industry that threatens our very existence." He then put his investigative skills to work. The result: more than 200 voice-over examples using his replicated voice. He was not aware they existed. He was never asked. He was never contacted. He gave no consent to the people or companies using his voice or voiceprint, and he notes that some of the apparent offenders appear to be located outside the United States.
What the lawsuits allege — and against whom
The legal response is already underway. Chicago-based attorney Ross Kimbarovsky of the firm Loevy & Loevy has filed 13 proposed class actions against major technology companies, including Google, Amazon, Meta, Nvidia, Microsoft, Apple and ElevenLabs, among others.
Kimbarovsky's accusation is sweeping and quotable: "What we are seeing is an illegal and unethical exploitation of talent on a massive scale, and one of the largest violations of biometric privacy ever committed. They've built a billion-dollar industry on stolen voices because they thought no one would make them pay for it."
The word "biometric" matters here. It frames a voice not as content but as a measurable personal identifier — the same category of data that fingerprint and face-recognition laws have long protected. That framing is the legal hinge on which the Illinois statutes, and potentially the federal bill, turn.
Which laws are being tested?
The class actions rest on two Illinois laws, according to Kurtis's account:
- One prohibits using AI to create and distribute "digital replicas" of an individual's voice, image or likeness — a digital replica being a synthetic recreation of a real person's identifiable traits.
- The other prohibits using audio without notice and permission to train an AI voice model capable of generating synthetic voices — striking directly at the data-collection practices of voice-cloning companies.
At the federal level, parallel legislation is moving. The No Fakes Act, supported by the entertainment union SAG-AFTRA, was approved by the U.S. Senate Judiciary Committee and now awaits a vote by the full Senate.
The stakes are considerable. Illinois has historically been the most litigated state in American biometric privacy law, and its statutes carry damages provisions that make class actions financially viable. If courts accept that a cloned voice is a "digital replica" under these laws, the exposure for companies that trained voice models on scraped audio — audio Kurtis says was taken "in most cases without compensation" — could be substantial.
Why Kurtis frames this as theft, not technology
Kurtis's article is not only a legal account. It is an argument about value, consent and the economics of creative work in the age of generative AI.
He describes data centers as appearing "like mushrooms after a rain. No announcement of their coming. No explanation." He compares their sudden emergence to "Stanley Kubrick's monolith in the desert," and warns that they "would promise to give us the world so they could take our jobs — our soul."
He extends the critique beyond his own case. The companies, he writes, have operated "like a giant vacuum cleaner," collecting knowledge from libraries, books, television and universities, then selling it back to subscribers for a fee. Their defense, he notes, is that this constitutes "fair use" — the copyright doctrine permitting limited unlicensed use of protected material — and therefore requires no compensation.
Kurtis identifies a common attitude linking the voice piracy he experienced and the bulk accumulation of training data: the belief that AI firms "have found an abundant, no-cost source to fuel their second industrial revolution. And they don't even have to tell anyone how they do it."
He also raises the physical footprint of the infrastructure itself, writing that AI is "starting to rise out of its silence as data centers strategically attempt to become operational before the public learns the full reach of the center's noise pollution, excessive water consumption and power demands."
What is at stake for a 60-year career
The core of Kurtis's argument is economic, and he states it in plain terms: "If many can get my voice for free, it soon will have no value for compensation. That can kill a career."
For Kurtis, the asset at risk took six decades to build. He describes his 60-year career as one "carefully crafted with integrity and ethical standards." The damage, in his view, is dual: the cloning steals not only his voice but also his journalistic reputation — the trust listeners place in a narrator they recognize. His closing position leaves no room for negotiation: "I'm not going to give it away."
That sentiment places Kurtis alongside the actors, narrators and voice artists represented by SAG-AFTRA, whose support helped push the No Fakes Act through the Senate Judiciary Committee. For working performers, the math is unforgiving: a synthetic replica that costs pennies to generate competes directly with the human original, whose rates reflect decades of craft.
What happens next
The immediate tests are procedural. The 13 proposed class actions filed by Kimbarovsky must survive motions to dismiss and win class certification before the Illinois statutes are applied to AI voice models at scale. In parallel, the No Fakes Act faces a full Senate vote that would create, for the first time, a federal baseline protecting voice, image and likeness from unauthorized AI replication.
Kurtis, for his part, has already done his own counting: almost 1,000 authorized narrations across 60 years, set against more than 200 unauthorized ones discovered after a single viewer's email. The outcome of the Illinois cases and the Senate vote will determine whether that ratio becomes a legal liability for the companies involved — or the new normal for everyone whose voice can be cloned.
Source: feed.koreatimes.co.kr
More from Marcus Bennett
Show full bio
Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
613 articles
Related articles
- Tokyo court rules AI-cloned voice violates actor's rights
- Modulate raises $25M to detect deepfake voices and read customer intent
- Modulate Cuts AI Voice Detection Time to 1.6 Seconds, Adds Twilio Ventures Backing
- Modulate Cuts AI Voice Detection to 1.6 Seconds, Adds Twilio Backing
- Over a Billion People Now Confide in AI Chatbots