From personal conversations to medical appointments, Apple’s Siri has allegedly been listening in on a wide range of private interactions through iPhones, HomePods, and Apple Watches, using the captured snippets to tailor ads on user devices. Although Apple has consistently denied trading intimate user details with advertisers, the tech giant recently agreed to pay $95 million to settle a lawsuit where plaintiffs noted disturbing instances—talking about Air Jordan sneakers and then immediately being bombarded with shoe ads; discussing Olive Garden and shortly thereafter being targeted with pasta ads; or having a confidential conversation with a doctor about surgery only to later see advertisements for the same procedure. In early May, a claims website was launched, inviting U.S. owners of any Siri-enabled device purchased between September 2014 and December 2024 to claim a compensation of up to $20 per device, a sum that might just cover a casual drink and a suspicious glance at your smartphone.

The case, Lopez v. Apple, initially came to light in July 2019, following a Guardian report based on revelations from an anonymous whistleblower, who was contracted by Apple to review Siri activations. This individual exposed how Siri often accidentally activated and recorded private audio due to background noises—sometimes as innocuous as a zipper sound. These recordings, disturbingly, included user location and contact details. Apple had not previously informed users that human reviewers might listen to their Siri interactions. Following the exposé, Apple suspended the review program within a week, and shortly thereafter, issued an apology, ensuring users that human review would be optional and discontinuing the default retention of audio recordings. This apology aimed to pacify user concerns rather than admitting any fault, and Apple denied all allegations from the lawsuit, a typical strategy in U.S. class-action settlements.

Supporting Quality Science Journalism

If you find value in this piece, consider supporting our award-winning journalism by subscribing. Your subscription helps sustain stories that delve into the significant discoveries and concepts that shape our modern world.

This scenario may seem familiar. Back in 2018, Amazon’s Alexa inadvertently recorded a private conversation about hardwood floors between a married couple and sent it to an employee of one of the spouses. Amazon attributed this breach to a series of misunderstood commands, essentially blaming it on a high-tech version of a pocket dial. A year later, Bloomberg revealed that Amazon employed thousands of people to transcribe voice clips to improve Alexa’s accuracy. Google later faced similar scrutiny. It became evident that ensuring the accuracy of voice-activated assistants required human training, which inadvertently led to privacy invasions through consumer devices, including televisions—Samsung even warned its smart TV users in 2015 not to discuss sensitive topics near their TVs because voice commands might be shared with third parties.

Such concerns are not just conspiracy theories. A 2019 survey indicated that over half of Americans believe their smartphones eavesdrop on them to gather data for ads, and a 2023 survey showed this belief had grown. In the UK, a 2021 survey showed that two-thirds of adults had seen ads they believed were triggered by their recent conversations. Psychologists attribute this phenomenon to a feedback loop enhanced by confirmation bias: we overlook the numerous irrelevant ads we encounter daily but fixate on the occasional ad that seems suspiciously timed after a private conversation. This has led to a pervasive cultural anxiety, prompting some to cover their device microphones with tape and others on TikTok to plead with Siri to respect their privacy. Given the data-driven nature of tech companies, such concerns are not unfounded.

Despite Apple’s previous advertising claim that “What happens on your iPhone, stays on your iPhone,” the recent settlement casts a shadow over its image. If even Apple can’t prevent these privacy breaches, then the question remains: who can? Apple, when contacted by Scientific American, reiterated its dedication to user privacy in its statement regarding the settlement. Likewise, Amazon reaffirmed its commitment to privacy, asserting that “Access to internal services is highly controlled, and is only granted to a limited number of employees who need these services to refine and enhance the service.” By the time this article was published, neither Samsung nor Google had responded to requests for comments.