"They've built a billion-dollar industry on stolen voices because they thought no one would make them pay for it," said a lawyer for the pla
In yet another display of how Illinois’ pioneering biometric privacy law can be used to protect Americans, state residents who work as audio storytellers, broadcast journalists, podcasters, voice actors, and more filed class-action lawsuits against Big Tech this week for “stealing their voices” to develop artificial intelligence products.
Since Illinois legislators passed the groundbreaking Biometric Information Privacy Act (BIPA) in 2008—regulating the collection, use, safeguarding, handling, storage, retention, and destruction of biometric identifiers, including fingerprints, voiceprints, and scans of a retina, iris, hand, or face geometry—there have been thousands of lawsuits filed and major settlements with Clearview AI, Facebook, and Six Flags.
Represented by the award-winning civil rights firm Loevy + Loevy, the Illinoisans are suing Adobe, Alphabet and its subsidiary Google, Apple, Amazon, ElevenLabs, Facebook parent company Meta, Microsoft, NVIDIA, and Samsung under BIPA.
The plaintiffs are audiobook narrators Lindsay Dorcus and Victoria Nassif as well as journalists Robin Amer, Yohance Lacour, Carol Marin, and Phil Rogers. Journalist Alison Flowers is part of all lawsuits except those against Amazon and Apple. Their lawyers noted that “between them, they have multiple Emmy and Peabody awards, several Pulitzer Prizes, several Alfred I. duPont-Columbia University awards, an Edward R. Murrow award, a James Beard award, a SOVAS award, and many, many other honors.”
Their cases focus on the voiceprint of each plaintiff, which is “a digital fingerprint of the human voice,” as the complaints explain. “It is a mathematical capture of the acoustic features—pitch, timbre, resonance—that emerge from a person’s distinctive physiology, combined with the speech patterns that person develops over a lifetime: accent, cadence, articulation. Like a fingerprint, a voiceprint identifies the individual. Like a fingerprint, it cannot be changed.”
The Adobe case targets Firefly, the company’s family of generative AI models. The complaint states that the company “treated the human voices that built Firefly as ownerless—ignoring the speakers’ rights, taking their voiceprints without asking, paying them nothing, and giving them no notice that their voices were being used at all, and ”built a mirage of commercial safety around products whose construction violated the one thing Illinois law requires before collecting a voiceprint: consent from the person.“
The Google filing points out that the company “has been a repeat defendant in BIPA cases” and even “paid approximately $100
million to settle BIPA claims arising from Google Photos’ face grouping feature,” among other high-profile settlements.
The Meta suit highlights that “no defendant in any biometric-privacy matter pending in the United States has had more direct, more sustained, or more financially consequential notice of BIPA than Meta,” given that the company “has paid the three largest biometric-privacy settlements in American history,” including $650 million to resolve claims under the Illinois law regarding Facebook’s photo tag suggestions.
“By the time Meta released Voicebox in June 2023, MMS in May 2023, and SeamlessM4T in August 2023, Meta had been a BIPA defendant for nearly a decade and had paid more than $2 billion in biometric-privacy settlements,” the complaint continues. “The technology Meta built using plaintiffs’ voices now competes with plaintiffs in the markets where they earn their living.”
The Amazon filing details similar harm to plaintiffs:
Amazon extracted plaintiffs’ voiceprints without notice or consent, depriving them of the right BIPA guarantees to make an informed decision about the collection and use of their biometric data. Amazon retains those voiceprints in its commercial models and continues to profit from them. Amazon has further disseminated those voiceprints, encoded in model parameters, through its cross-affiliate, subprocessor, and integration-partner networks. The technology built on those voiceprints now displaces plaintiffs in the markets where they earn their living—the broadcast journalism, investigative podcast, audiobook narration, voiceover, and voice performance markets that the voice products are designed and sold to serve.
“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,” said Loevy + Loevy attorney Ross Kimbarovsky in a Thursday statement.
“The legislators who wrote and passed BIPA had the foresight to realize that biometric privacy was going to be a major civil rights issue in the 21st century,” the attorney continued. “Social security numbers can be changed, passwords can be reset, and credit cards can be canceled, but once your biometric data is compromised, there’s nothing you can do about it.”
“These companies know the law, know their liability, and know exactly how to build consent systems that comply with BIPA,” Kimbarovsky added. “They’ve built a billion-dollar industry on stolen voices because they thought no one would make them pay for it.”
In addition to Illinois, Texas and Washington state have enacted biometric privacy laws, while California, Colorado, Connecticut, Utah, and Virginia have comprehensive consumer protection policies that apply to such information, according to Bloomberg Law. However, efforts in Congress to enact federal legislation—such as the National Biometric Information Privacy Act and the Facial Recognition and Biometric Technology Moratorium Act—have been unsuccessful.
I love technology and it's advances, but the moment you mention AI or biometrics I'll turn straight into Dale Gribble. No you may not have my fingerprint to log into my facebook. Fuck. Them. Clankers.
Cap_able has launched a collection of knitted clothing that protects the wearer's biometric data without the need to cover their face.
"Named Manifesto Collection, the clothing features various patterns developed by artificial intelligence (AI) algorithms to shield the wearer's facial identity and instead identify them as animals."
"The collection includes jumpers, t-shirts, trousers and dresses that are knitted from 100 per cent Egyptian cotton quality by Filmar, a brand that adheres to the Better Cotton Initiative (BCI), according to Cap_able.
According to Didero, the Manifesto Collection aims to raise awareness of the right to biometric data privacy, which is an issue she says that is "often underrepresented despite affecting the majority of citizens around the world"."
The most recent edition of Shed Letters is available!
I fell down a rabbit hole about the ethical, legal, and existential issues (and potential!) of smart toilets and now I'm making it everyone's problem! Please I stayed up too late researching this. Please read it.
(the pic of Scribble is there because I start almost all my newsletters with pictures of my cats. Because obviously why wouldn't I?)
The startup promises a fairly-distributed, cryptocurrency-based universal basic income. So far all it's done is build a biometric database from the bodies of the poor.
Pete Howson, a senior lecturer at Northumbria University who researches cryptocurrency in international development, categorizes Worldcoin’s actions as a sort of crypto-colonialism, where “blockchain and cryptocurrency experiments are being imposed on vulnerable communities essentially because…these people can’t push back,” he told MIT Technology Review in an email.
What makes the crypto version even more harmful than other forms of data colonialism is that decentralization, the core tenet of blockchain, makes for “very limited accountability…when things go wrong,” Howson explained. “You’ll often hear this phrase ‘Do Your Own Research’, or DYOR, because these guys don’t care much for rules and regulations.”
But inequities in information and internet access make that “do your own research” ethos all but impractical for many people in developing regions. Similarly, huge economic disparity means that in Kenya, say, the promise of just under half a US dollar could be a compelling incentive for someone to give up their biometric data, whereas in Norway or the US, such an offer wouldn’t go far.
Simply put, it’s just cheaper and easier to run this kind of data collection operation in places where people have little money and few legal protections.
-
-
-
Speaking to Blania clarified something we had struggled to make sense of: how a company could speak so passionately about its privacy-protecting protocols while clearly violating the privacy of so many. Our interview helped us see that, for Worldcoin, these legions of test users were not, for the most part, its intended end users. Rather, their eyes, bodies, and very patterns of life were simply grist for Worldcoin’s neural networks. The lower-level orb operators, meanwhile, were paid pennies to feed the algorithm, often grappling privately with their own moral qualms. The massive effort to teach Worldcoin’s AI to recognize who or what was human was, ironically, dehumanizing to those involved.
When we put seven pages of reporting findings and questions to Worldcoin, the company’s response was that nearly everything negative that we uncovered were simply “isolated incident[s]” that ultimately wouldn’t matter anyway, because the next (public) iteration would be better. “We believe that rights to privacy and anonymity are fundamental, which is why, within the next few weeks, everyone signing up for Worldcoin will be able to do so without sharing any of their biometric data with us,” the company wrote. That nearly half a million people had already been subject to their testing seemed of little import.
Rather, what really matters are the results: that Worldcoin will have an attractive user number to bolster its sales pitch as Web3’s preferred identity solution. And whenever the real, monetizable products—whether it’s the orbs, the Web3 passport, the currency itself, or all of the above—launch for its intended users, everything will be ready, with no messy signs of the labor or the human body parts behind it.
The whole article is quite good and detailed and does not require deep knowledge of the subject of blockchain. It’s well explained in context here.
