AI-powered communications platform provider Dialpad is facing a class action lawsuit in California for allegedly eavesdropping on phone calls made by consumers to businesses using its services. The lawsuit, filed by San Francisco resident Gary Freund on behalf of a proposed class of California consumers, accuses Dialpad of secretly listening to and analyzing consumer conversations through its artificial intelligence tools – without consumer knowledge or consent – and then using that data to train its AI models for commercial purposes.
The complaint, filed in San Francisco County Superior Court, asserts that Dialpad violates California’s Invasion of Privacy Act, specifically Penal Code § 631, which prohibits wiretapping or reading the content of communications without consent from all parties. Freund, who contacted Warby Parker’s customer support line, alleges that Dialpad acted as an undisclosed third party during the call. According to the complaint, Dialpad not only transcribed and analyzed the conversation using AI in real time, but also retained portions of it to improve its machine learning algorithms – effectively using his data and that of other consumers to enhance its products for other clients.
Dialpad’s core product is a cloud-based communications suite that includes voice, messaging, and call center functionality for businesses. Embedded within this offering is “Dialpad AI,” which boasts capabilities like real-time transcription, sentiment analysis, and automatic call summaries. According to Dialpad, its AI system is built on a speech recognition model trained on more than five billion minutes of audio and message data. This data, as described in Dialpad’s own materials, includes “segments of call audio, video, or transcriptions” saved to improve its language models – meaning calls made to one Dialpad client are potentially used to improve service for others.
The lawsuit raises significant questions about consumer privacy in an era of increasingly pervasive AI technologies. Unlike traditional recording practices, where consumers are often notified that “this call may be recorded for quality assurance,” or other similar announcements, the complaint alleges that Dialpad is not merely recording, but actively analyzing, extracting sentiment, and compiling summaries – all while serving as a silent participant in conversations consumers thought were private. Worse, the complaint argues that this analysis isn’t just to assist the business receiving the call, but to benefit Dialpad’s broader customer base through AI model training.
Dialpad’s privacy policy, according to the filing, discloses that call data may be used for improving AI systems, but the plaintiffs argue that this falls far short of the explicit, informed consent required under California law. They also contend that consumers are unaware that a third-party AI service is involved at all, let alone using their voice data for separate commercial objectives. The suit claims this behavior amounts to illegal wiretapping and seeks statutory damages of $5,000 per violation, meaning every individual call could potentially yield significant financial liability for the company.
Freund’s case is being brought on behalf of all California residents who made or received phone calls that were analyzed by Dialpad while communicating with a Dialpad business customer. The plaintiff’s legal team is seeking class certification, an injunction to stop the alleged practice, and both compensatory and punitive damages.
The filing notes that Dialpad works with more than 70,000 enterprise customers, including prominent companies such as Netflix, Johns Hopkins, Cigna, T-Mobile, and Warby Parker, making the potential scope of the class action substantial. Given that many of these businesses may also be unaware of how Dialpad uses customer call data beyond the original service context, the lawsuit could ripple through the broader B2B software ecosystem. At issue is not only Dialpad’s conduct, but the entire practice of using real-world customer data to enhance artificial intelligence systems without full transparency.
This case also highlights the growing legal and ethical tension between AI innovation and privacy rights. As AI tools become more embedded in daily business functions – from contact center operations to sales calls and internal meetings – the question of who owns and controls the underlying data becomes increasingly urgent. California, which enforces some of the strictest privacy protections in the U.S., may become a bellwether for how courts interpret AI’s role in consumer communication and data use.
If the class action proceeds, it could compel major changes in how AI-powered communication platforms disclose their data practices. Companies may be required to provide clearer notice, secure explicit opt-in consent, or even segregate call data used for model training from the data used to assist with live conversations. Failure to comply could not only invite legal risk but also damage consumer trust at a time when scrutiny of AI systems is rapidly intensifying.
In addition, with so many different groups watching this case closely, the outcome could influence future legislation. Federal regulators and lawmakers are already considering stronger AI governance frameworks, and real-world examples like this one may provide fuel for additional guardrails around the use of AI in consumer-facing interactions.
Dialpad has not yet issued a public response to the complaint, but the stakes are high. A ruling against the company could set precedent and affect a wide range of vendors in the AI-driven communications and customer experience space. From startups to established tech giants, many rely on real-world conversational data to refine their products. How this case unfolds may determine how – indeed, if – that data can be used going forward.
The plaintiff’s core argument is simple but powerful: A consumer should not have to wonder whether a third-party AI system is analyzing their voice, extracting their tone, and using their words to benefit other companies. Rather, they should be informed, and they should be asked. Anything less, the lawsuit suggests, isn’t just poor ethics; it may be unlawful.
As AI technologies continue to proliferate across business communications, this case could define the boundaries of consent and surveillance in the age of machine learning. Whether in customer support, healthcare, finance, or retail, businesses that integrate AI into their workflows must balance innovation with transparency and, increasingly, legal compliance. The Dialpad case may be the clearest test of that principle yet and has the potential to send ripple effects across the AI landscape.
Edited by
Erik Linask
