When Echoes Resonate Too Closely: What Happens When AI Mimics a Human Voice?

Voices are the subtle architecture of human connection. They carry not just words, but cadence, nuance, and the gentle inflections that tell listeners who is speaking and how seriously to take them. For those in broadcasting, a well-crafted voice becomes both tool and signature — honed over years like a musician tuning an instrument to harmony.

That is the feeling behind a lawsuit now unfolding in California, one that blurs the boundary between human expression and the synthetic echoes of artificial intelligence. David Greene, a veteran public radio host known for his work on NPR’s “Morning Edition” and the political roundtable podcast “Left, Right & Center,” says that an AI voice used in a Google product sounds so much like his own that he worries it could damage both his reputation and the integrity of his craft.

Greene’s career was shaped over decades of practice, encouragement, and careful listening. Growing up in Pittsburgh, he would mimic baseball announcers and record school morning announcements long before he landed his first job with NPR in 2005. Over the years, he learned to temper his natural patterns — those “uhhs” and slurred intonations — into a voice that was warm, steady, and trusted by millions of listeners.

When friends and colleagues first asked Greene if he had worked with a new Google tool called NotebookLM, which lets users generate AI-driven podcasts, he listened out of curiosity — and then surprise. The male voice in the tool’s Audio Overviews, he says, sounded eerily like his own. “It’s this eerie moment where you feel like you’re listening to yourself,” he told reporters.

The resemblance, which he and his lawyers say goes beyond casual similarity in tone and cadence, has led Greene to file a lawsuit in Santa Clara County Superior Court. He alleges that Google used voice data that replicates his vocal identity without permission or compensation — a claim Google denies, saying the voice is based on a paid professional actor.

The case highlights a broader tension in the age of generative AI, where speech, image, and text can be synthesized at the push of a button. Legal experts point out that precedent exists for unauthorized imitation — for example, the successful 1988 lawsuit by singer and actress Bette Midler against a car company for using a sound-alike voice actor in a commercial — but the technology today raises fresh questions about consent, likeness rights, and technological stewardship.

For Greene, the concern is deeply personal as much as professional: “My voice is, like, the most important part of who I am,” he said. The fear is not simply of lost income, but of an AI-generated voice making statements or engaging in content that he would never endorse, yet listeners might assume comes from him.

In the background of this dispute are emerging legislative efforts in several U.S. states and at the federal level to regulate the use of people’s voices in AI tools. But Greene’s lawsuit relies on existing state laws granting public figures certain rights to control the commercial use of their likeness. Experts note that the outcome could help clarify how these laws apply in the rapidly evolving world of AI.

Observers both inside and outside the tech community say cases like this are likely to become more common as tools for generating lifelike speech proliferate. The legal questions center on how closely an AI voice must resemble a real person’s to constitute infringement, and whether ordinary listeners would mistake one for the other.

In simple terms, David Greene, a longtime radio host, is suing Google, alleging that an AI voice in its NotebookLM product so closely resembles his own that it effectively “stole” his vocal identity without consent — a claim the company disputes.

The case will proceed in court, where questions of likeness, consent, and emerging norms for AI tool creation will be tested in an era where the human voice can be digitally reconstructed at scale.

AI Image Disclaimer Visuals are created with AI tools and are not real photographs.

Sources: The Washington Post Yahoo News Singapore (syndicated) TechCrunch Reuters The Guardian