Nine major technology companies, including Apple, Amazon, Meta Platforms, Microsoft, Nvidia, and Samsung, are facing lawsuits in Chicago federal court for allegedly using recorded human voices to train artificial intelligence without permission. The proposed class action lawsuits, filed by a group of journalists, podcasters, and audiobook narrators, allege the companies violated the Illinois Biometric Information Privacy Act, known as BIPA. The plaintiffs claim the companies harvested data from various sources to power their AI systems without obtaining the required consent or providing details on how the data would be used and destroyed.
BIPA is a state law that requires companies to inform individuals and obtain written consent before collecting biometric data. It also mandates that companies publish written policies regarding the retention and destruction of such information. While Illinois legislators narrowed the law's penalties in 2024, companies can still be fined $1,000 for each negligent violation or $5,000 for each intentional or reckless violation. The plaintiffs in these cases include Pulitzer Prize winners Yohance Lacour and Alison Flowers, as well as Chicago journalist Carol Marin, who allege violations of BIPA and state laws concerning consumer fraud and publicity.
In response, the tech companies have filed motions to dismiss the claims, arguing that the lawsuits are speculative and that the Illinois law does not apply because the alleged data misuse did not occur within the state. Apple's attorneys stated in a filing on Monday that the plaintiffs only showed their recordings were available on the internet but provided no evidence the company used them for AI training. A Meta spokesperson said the company provides users with information about data usage and characterized the claims as incorrect. The cases are currently distributed among seven different judges, though Apple has requested they be consolidated.
A person participating in these creative industries would notice the impact through legal precedents establishing whether their vocal characteristics are protected biometric data similar to fingerprints or facial scans. If the plaintiffs prevail, it would require tech companies to change their day-to-day data collection practices by implementing formal consent forms and data-destruction schedules for voice recordings. Conversely, a victory for the tech companies would affirm their argument that state-level privacy laws do not extend to data harvested from the general internet for AI development.
The litigation also sets a precedent for how biometric laws interact with the growing AI market, potentially influencing future policy and how other states regulate data scraping. The next steps in the legal process include rulings on the companies' motions to dismiss and a decision on Apple's request to consolidate the nine cases. While a specific trial date has not been reported, the recent filings by companies like Apple and Meta indicate the defense will focus on jurisdictional arguments regarding where the data processing actually occurred.