Article Summary
- A group of well-known Chicago journalists, podcasters and voice actors are behind nine class-action lawsuits filed this week alleging major tech companies used their voices without their permission in order to train AI products.
- The lawsuits represent a new area of focus for Illinois’ strongest-in-the-nation biometric data privacy law, which has spawned thousands of lawsuits in the last decade or so. The litigation has netted Illinoisans millions of dollars in settlements, mostly over the collection of employees’ fingerprints in timeclock technology.
- Plaintiffs include locally famous broadcast journalists Carol Marin and Phil Rogers, both retired from Chicago’s NBC 5 news station, along with prolific audiobook narrators and podcasters.
This summary was written by the reporters and editors who worked on this story.
CHICAGO — Over hundreds of pages in legal filings this week, a group of well-known Chicago-based journalists, podcasters and voice actors accused tech giants like Google, Amazon, Apple, Microsoft and others of “stealing” their voices to train Artificial intelligence.
The nine class action lawsuits, filed in Chicago’s federal court between Monday and Wednesday, represent a new frontier for Illinois’ strongest-in-the-nation biometric data privacy law. In the last decade or so, the state’s Biometric Information Privacy Act, or BIPA, has spawned thousands of lawsuits against companies alleged to have collected and stored biometric data from employees and customers without proper notice or consent.
The vast majority of that litigation — which has paid out millions of dollars to Illinoisans mostly via class-action settlements — has been over employee fingerprints collected by timeclock technology, though Facebook’s $650 million settlement in 2020 was with users over facial recognition.
Read more: Court rulings supercharge Illinois’ strongest-in-nation biometric privacy law
But as companies adopted policies to comply with BIPA and the pool of plaintiffs began to dry up, technology has rapidly evolved. Smart security cameras, safety cameras focused on workplaces, online “try-on” technology that allows users to envision, for example, what a certain pair of glasses would look like on their face, have become popular targets of BIPA litigation.
And with the breakneck speed of AI development, companies focused on building that technology could prove to be the next major focus for BIPA lawyers.
In the cases filed this week, locally famous broadcast journalists like Carol Marin and Phil Rogers, both retired from Chicago’s NBC 5 news station, along with podcast hosts and voice actors allege the companies ingested recordings of their voices in order to train their AI “foundational voice models.”
“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,” Ross Kimbarovsky, an attorney with Chicago-based law firm Loevy & Loevy, said Thursday in a statement announcing the lawsuits.
Kimbarovsky accused the companies of disregarding BIPA despite knowing “exactly how to build consent systems that comply with BIPA.”
“They’ve built a billion-dollar industry on stolen voices because they thought no one would make them pay for it,” he said.
Other plaintiffs include journalist Robin Amer, audiobook narrators and voice actors Lindsay Dorcus and Victoria Nassif, and podcasters Yohance Lacour and Alison Flowers — all Illinois residents.
Tech heavyweights named
The lawsuits name Amazon, Adobe, Google and its parent company Alphabet, Apple, Microsoft and Samsung, as well as Facebook parent company Meta, text-to-speech AI company ElevenLabs, and advanced computer chip maker NVIDIA. None of the companies responded to a request for comment on the lawsuits.
BIPA defenders point out that biometric information is unique, and losing control of it can be irrevocable. If an individual’s Social Security number is stolen, for example, it may be a nuisance to get a new one but not impossible. But there’s no remedy for a stolen fingerprint, retinal, voice or face scan, they argue. Under the law, companies deploying this technology must obtain written consent before biometric information is collected.
But the lawsuits allege the companies never gave anyone a chance to give consent for their voiceprints to be ingested into their AI training models.
“None of them was told that their voice was being used to train Amazon’s commercial voice AI,” the lawsuit against Amazon said. “None of them was asked. None of them consented.”
A voiceprint “is a digital fingerprint of the human voice,” according to the complaints, which go on to characterize it as “a mathematical representation” of someone’s voice, including pitch, timbre and resonance determined by a speaker’s physiology. A voice is also defined by speech patterns “developed over a lifetime,” including accent, cadence and articulation.
“Like a fingerprint, a voiceprint identifies the individual and cannot be changed,” the lawsuits say. “A Social Security number can be reissued. … A person whose voiceprint has been taken cannot recover it by altering their voice — the biological and behavioral patterns that produced the voiceprint are the same ones used to speak every day.”
Voiceprint-focused lawsuits may very well become fertile ground for BIPA, especially if judges weighing the complaints filed this week agree the cases should move forward. Industry experts believe the cases could hinge on whether the voiceprints are identifiable.
In early 2023, upscale grocer Whole Foods — which was acquired by Amazon in 2017 — settled a case brought by 330 warehouse employees who alleged the company collected their voiceprints without permission and used them to verify workers’ identities. The $300,000 payout was the first BIPA settlement resulting from voiceprint-focused litigation.
Capitol News Illinois is a nonprofit, nonpartisan news service that distributes state government coverage to hundreds of news outlets statewide. It is funded primarily by the Illinois Press Foundation and the Robert R. McCormick Foundation.
