Yes, journalists’ recordings were allegedly used without permission to train Google’s AI voice models. In May 2026, award-winning journalists including Chicago-based reporter Carol Marin, along with Pulitzer Prize winners Yohance Lacour and Alison Flowers, filed a class-action lawsuit in Chicago federal court alleging that Google, Meta, Microsoft, Nvidia, and ElevenLabs scraped their long-form, studio-quality audio recordings to train AI models including Google Gemini Live and Google Assistant. The lawsuit specifically names Carol Marin, whose distinctive voice and decades of journalism made her recordings particularly valuable for training AI voice systems.
The nine simultaneously filed lawsuits center on alleged violations of Illinois’s Biometric Information Privacy Act (BIPA), the nation’s strictest biometric data privacy law. According to court documents and reporting by PYMNTS.com and the Chicago Sun-Times, the tech companies targeted “long-form, single-speaker, studio-quality” recordings that matched their documented criteria for optimal AI training audio. The journalists claim they never consented to having their voices scraped, analyzed, or used to train commercial AI products—a violation that could entitle them and other affected creators to statutory damages under BIPA’s provisions.
Table of Contents
- How Did Tech Companies Allegedly Obtain Journalists’ Recordings for AI Training?
- Understanding the Biometric Privacy Violation Under Illinois BIPA
- Who Are the Named Plaintiffs and What Makes This Case Significant?
- What Are Your Legal Rights and How Can You File a Claim?
- What Challenges and Defenses Might the Tech Companies Raise?
- The Broader Impact on Content Creators and Media Professionals
- What This Case Means for the Future of AI Voice Training and Content Rights
- Conclusion
How Did Tech Companies Allegedly Obtain Journalists’ Recordings for AI Training?
The lawsuit alleges that the defendants systematically scraped recordings from publicly available sources, including podcasts, news broadcasts, and audio archives featuring the named plaintiffs. According to PYMNTS reporting, Google and its co-defendants specifically targeted “long-form, single-speaker, studio-quality” content—exactly the type of professional-grade audio that journalists and narrators produce. Rather than licensing or purchasing the rights to use these recordings, the companies allegedly harvested them from the internet without the creators’ knowledge or permission.
The recordings in question represent decades of professional work. Carol Marin, a legendary Chicago journalist known for her investigative reporting and distinctive broadcast presence, had her voice and likeness embedded in these AI models. The defendants’ algorithms identified and downloaded high-quality audio files featuring the plaintiffs, then used machine learning techniques to extract vocal biometric data—essentially creating a digital replica of each person’s voice. This extraction process violates BIPA because voice is classified as biometric information under Illinois law, and the law requires explicit written consent before any company collects, uses, or profits from someone’s biometric data.
Understanding the Biometric Privacy Violation Under Illinois BIPA
Illinois’s Biometric Information Privacy Act stands apart from privacy laws in other states because it grants individuals statutory damages of $1,000 to $5,000 per violation, making it one of the most powerful consumer protection tools in the nation. A “violation” can mean a single unauthorized collection, use, or storage of biometric information. When applied to AI training, each time a defendant’s algorithm processed a journalist’s voice recording to extract biometric data, that could constitute a separate violation. This means damages could accumulate rapidly, especially if the recordings were processed multiple times during model development.
However, a key limitation in the lawsuit’s success will be proving that the defendants actually extracted and used voice biometric data for training, rather than simply downloading audio files. The tech companies will likely argue that they only used the audio as training material without explicitly targeting biometric extraction. Additionally, some defendants may claim they obtained the recordings from legitimate third-party sources that already held the rights. The lawsuit’s outcome will likely hinge on forensic evidence showing exactly what the companies’ machine learning models extracted from the audio and how they used that data.
Who Are the Named Plaintiffs and What Makes This Case Significant?
The named plaintiffs represent a high-profile mix of journalism and media talent. Carol Marin, the lead Chicago-based plaintiff, is an Emmy Award-winning journalist and founder of Marin Investigates, with decades of on-air presence and credibility. Yohance Lacour, a Pulitzer Prize winner, and Alison Flowers, another Pulitzer Prize-winning journalist, round out the named class representatives. Their prominence makes this case impossible for the defendants to ignore—these are not unknown voices but recognizable public figures whose voices hold significant commercial value in the AI training marketplace.
The plaintiffs’ high-profile status also underscores a broader issue: if major tech companies were willing to scrape the recordings of award-winning journalists, they almost certainly harvested content from less prominent creators as well. Podcasters, audiobook narrators, radio hosts, and other voice professionals likely had their work used without consent. The class action structure means that other affected journalists, narrators, and voice professionals can join the case and seek damages. The lawsuit doesn’t require that you be as famous as Carol Marin or Yohance Lacour to have standing—anyone whose recorded voice was harvested and used to train these AI models could potentially qualify as a class member.
What Are Your Legal Rights and How Can You File a Claim?
If your voice recordings were used to train these AI models without your consent, you have potential rights under BIPA and similar state biometric privacy laws. The class-action structure means you don’t need to hire your own lawyer or prove individual damages—the plaintiffs’ attorneys handle the litigation, and class members can recover a share of any settlement or judgment. To participate, you’ll typically need to file a claim with the settlement administrator once a settlement is reached, providing evidence that your recordings were used (such as links to podcasts, news archives, or other sources where your voice was publicly available).
One important tradeoff: settling a class-action lawsuit usually requires you to accept the settlement amount in exchange for releasing your right to sue the defendants individually. However, given that proving individual damages under BIPA requires extensive litigation, joining the class action is often more practical than pursuing your own case. You should monitor the lawsuit’s progress through legal databases and news outlets, as the settlement administrator will eventually announce claim filing deadlines. Missing a filing deadline typically means forfeiting your right to recover, so vigilance is essential.
What Challenges and Defenses Might the Tech Companies Raise?
The defendants will likely argue several points in their defense. First, they may claim that downloading publicly available audio files doesn’t require consent under fair use or the Computer Fraud and Abuse Act exceptions. Second, they could argue that they used the audio for legitimate research or product development and didn’t explicitly extract or misuse voice biometric data. Third, some defendants (particularly Microsoft and Nvidia) may assert they didn’t directly scrape the content but licensed or received it from intermediaries, shifting liability away from themselves.
A significant limitation in this lawsuit is proving the specific intent and mechanics of the AI training process. Tech companies maintain that their machine learning systems are complex “black boxes”—even their engineers may not be able to trace exactly what data was extracted and how it was used. This opacity makes it harder for plaintiffs to prove that voice biometric data was deliberately extracted versus incidentally processed. Additionally, the defendants will argue that Illinois BIPA’s privacy protections don’t extend to recordings obtained from public sources, a legal argument that could narrow the scope of liability if courts accept it. The litigation timeline could stretch years, with appellate battles likely regardless of the initial outcome.
The Broader Impact on Content Creators and Media Professionals
This lawsuit sends a warning signal to all voice-based content creators: your recordings have commercial value to AI companies, and that value is being extracted without your knowledge or compensation. Podcasters, journalists, narrators, voice actors, and radio professionals have built their careers on the quality and uniqueness of their voices. The allegation that AI companies harvested this work to build competing commercial products represents a fundamental breach of creator rights. The case highlights how the rapid advancement of AI has outpaced legal protections for individual creators.
The litigation could reshape how media professionals approach distribution and licensing. Some creators may choose to publish only behind paywalls or restricted platforms to limit scraping. Others may demand upfront consent and licensing fees before allowing their work to be used for AI training. The outcome of this case could establish precedent that forces AI companies to negotiate with content creators rather than simply harvesting content from the internet. For now, the message is clear: if your voice is distinctive and recognizable, it has value in the AI marketplace—and you have the right to control whether that value is extracted and monetized.
What This Case Means for the Future of AI Voice Training and Content Rights
This lawsuit represents one of the first major legal challenges to the broad-scale scraping practices that have underpinned AI training datasets. As more creators realize their work was used without permission, similar lawsuits will likely follow against other AI companies and other categories of content creators. The outcome could establish a new legal framework where companies must obtain consent and potentially pay licensing fees to use creative work for AI training—fundamentally changing the economics of AI development.
Looking forward, this case could accelerate the shift toward consensual AI training partnerships, where creators are informed their work will be used and receive compensation. Alternatively, if the defendants prevail, it could entrench the practice of unrestricted scraping, creating an arms race where AI companies move fastest because they don’t have to negotiate with creators. The stakes extend beyond journalists to all digital creators: musicians, artists, writers, and software developers whose work could be similarly harvested for AI training.
Conclusion
Journalists’ voices were allegedly scraped without consent to train Google Gemini Live, Google Assistant, and other AI models developed by Google, Meta, Microsoft, Nvidia, and ElevenLabs. Nine class-action lawsuits filed in May 2026 by award-winning journalists including Carol Marin, Yohance Lacour, and Alison Flowers allege violations of Illinois’s Biometric Information Privacy Act. The case centers on the companies’ alleged systematic harvesting of “long-form, single-speaker, studio-quality” recordings—precisely the type of professional audio that training algorithms prioritize. If the plaintiffs succeed, they could secure statutory damages under BIPA’s powerful penalty provisions and establish that voice biometric data requires explicit consent before use.
If your recorded voice was used to train these AI models without your permission, you have potential rights as a class member. Monitor the lawsuit’s progress through legal news outlets and be prepared to file a claim once a settlement is reached and the administrator announces claim periods. The broader implication is that content creators of all types—journalists, podcasters, narrators, and voice professionals—now understand their recordings have significant commercial value. This lawsuit could force AI companies to negotiate fair licensing agreements instead of simply harvesting creative work from the internet, ultimately reshaping how AI training datasets are built and who profits from them.
You Might Also Like
- Wiley Rein Data Breach Class Action Claims Hackers Stole Sensitive Client Information
- Takeda Amitiza Antitrust Class Action Claims Generic Competition Was Improperly Delayed
- Skechers Email Marketing Class Action Claims Sale Messages Created False Urgency
Open Settlements You Can Claim Now
Browse current class action settlements accepting claims — several require no proof of purchase: