Following the Zaluda ruling — where a Cook County Circuit Court judge granted class certification to millions of Illinois residents over Apple Siri's unauthorized voice data collection — five Illinois residents sued Microsoft in February, alleging that the Teams transcription feature creates voiceprints without BIPA-compliant consent.

The class action lawsuit, Basich et al. v. Microsoft Corporation (Case No. 2:26-cv-00422), was officially filed on February 5, 2026, in the U.S. District Court for the Western District of Washington by five Illinois residents — Alex Basich, Kristin Bondlow, Marquis Boyce, Jessica Brewer, and Jamari Brown — all of whom participated in Teams meetings where live transcription was enabled.

The complaint targets Microsoft Teams' real-time transcription feature, alleging that Teams utilizes a machine learning process called "speaker diarization." Speaker diarization is the AI-driven process of partitioning an audio or video recording into segments based on speaker identity. Simply put, it answers the question "who spoke when?" and is used to generate clean, speaker-labeled transcripts rather than a raw, unformatted block of text.

The moment a system distinguishes one voice from another to attribute speech in a transcript, the plaintiffs argue, it has crossed into biometric territory. Whether the meeting host enabled transcription, whether participants downloaded the application, whether the voiceprint was retained or discarded, all of it becomes secondary. The plaintiffs argue this process forces the creation of mathematical voiceprints (biometric identifiers) without BIPA-compliant written disclosures, public data retention timelines, or explicit written consent. For this violation, they ask the Court to grant statutory damages ($1,000 for negligent violations; $5,000 for intentional/reckless violations).

The surge in self-represented litigants leveraging generative AI to draft and file court papers is a trend federal judges observe with increasing worry. Nippon’s lawsuit is among the first cases to explicitly charge a major AI platform with the unauthorized practice of law – as such, it represents a critical stress test for the bar and the bench alike. 

Nippon Life Insurance Company of America v. OpenAI, filed March 4 in the Northern District of Illinois is something out of a Netflix show. The factual premise is almost surreal. It centers around a long-term-disability claim from a former employee of Nippon Insurance. Nippon Life Insurance is suing OpenAI on three claims: 

What makes this case important is that Nippon Life uses OpenAI's October 2025 update to its terms of service, which prohibited tailored legal advice, as evidence that OpenAI itself recognized the foreseeable risk. 

In a Friday filing in Chicago federal court, OpenAI countered that there are no legitimate grounds for the lawsuit brought by Nippon Life Insurance Company.

On April 29, families affected by the Tumbler Ridge, British Columbia school shooting filed suit against OpenAI in California state court, alleging that the company was negligent for failing to alert authorities after the shooter's account was internally flagged for "gun violence activity and planning" months before the attack. The complaint alleges OpenAI's safety team recommended notifying authorities, that company leadership instead deactivated the account, and that the shooter created a second account and continued. CEO Sam Altman issued a public apology days before the filing.

These cases are following a defective-product framework that has been used against social-media platforms for several years and is now being adapted for AI chatbots. For practitioners advising AI companies, the architectural and governance question raised by both Nippon Life and the Tumbler Ridge filings is the same — what does a reasonable safeguard look like, and at what point does the failure to implement it become foreseeable harm? For AI companies, this underscores that establishing internal safety protocols is no longer just an ethics exercise.