Consumers who own smart home devices, such as Amazon’s Echo smart speaker, are the most likely to feel that their device is listening unprompted, with seven in 10 saying they felt it was either very or fairly likely to be eavesdropping on them.
Smartphone users had similar concerns, with three in five believing it’s fairly (30 per cent) or very (31 per cent) probable their handset was actively listening, according to YouGov research of 2,157 adults.
The freshest exclusives and sharpest analysis, curated for your inbox
Rumours and theories have abounded for years that tech companies are eavesdropping on their users via devices’ microphones in order to better target adverts for products and services.
While two-thirds (66 per cent) of adults surveyed said they’d been targeted with adverts after speaking about a particular product, one fifth said they suspected this was informed by their device eavesdropping on them.
Dr Max Eiza, lecturer in computer and network security at UCLan, said smartphones are in continuous listening mode waiting for a keyword to wake the digital assistants that have become commonplace in handsets, including Apple’s Siri, Amazon’s Alexa and Google Assistant.
“There is no guarantee that the conversation before that keyword is triggered is not recorded,” he told i.
“It’s happened to me: when I’ve been talking about something, then all of a sudden when I look at the feed on Google I see products coming through or news related to what I was talking about.
“The companies can deny it as much as they like, but there’s a lot of evidence,” he added. “Unless a whistleblower comes out and says something, you have to ask yourself what kind of scandal it would take in order for people to stop using a product. If you want to be sure a device is not listening, turn it off.”
A 2019 study from mobile security company Wandera set out to test whether an iPhone and Samsung Galaxy smartphone displayed adverts for pet food after being exposed to a pet food video playing on loop for 30 minutes over three days.
It found no evidence of pet food adverts following the experiment, suggesting that the tracking techniques advertisers deploy to follow users across the web and tailor adverts to their needs are so sophisticated that “advertisers don’t need to listen to our conversations”.
However, it pointed out that just because advertisers don’t appear to be listening in through devices’ microphones, that doesn’t necessarily mean that no one is, adding that hackers regularly attempt to take control of microphones and intercept audio files.
“If consumers wish to continue to pay the lowest price possible for smart devices that might listen in and offer services or products based upon intercepted conversations, messages or usage, then that is a trade-off they must be willing to make”, Thom Langford, Security Advocate at security firm SentinelOne, said.
“The unfortunate side of this, however, is that very rarely do these manufacturers make it overtly clear that this trade-off is taking place, or if they do, it is buried deep in the small print of the multi-page End User License Agreement (EULA).
“Governmental regulation forcing transparency may be the only approach to ensuring consumers make informed decisions around their digital privacy.”
While Big Tech has long denied that it snoops on its customers to personalise adverts, the internet is awash with instances of users convinced it’s the only way their device would know to present them with adverts.
Facebook issued a statement in 2016 stating that it did not use a phone’s microphone to inform ads or to influence what appeared in a user’s news feed.
“Some recent articles have suggested that we must be listening to people’s conversations in order to show them relevant ads. This is not true,” the company said.
“We show ads based on people’s interests and other profile information – not what you’re talking out loud about.”
Similarly, Google has said it does not “use ambient sound from any device to target ads”.
Amazon has repeatedly denied its Alexa virtual assistant is always listening to user conversation, reiterating it only begins recording audio once it hears its default “Alexa wake word”, but was subject to considerable backlash when it emerged the company had hired thousands of human workers to listen to Alexa-captured voice recordings without making it explicitly clear they’d done so in 2019.
Apple, Facebook, Google and Microsoft all suspended their own use of human annotation contractors following the public backlash, with Apple and Google changing their permissions to an ‘opt-in’ model to allow humans to listen to recordings, while Amazon and Facebook adopted an opt-out position.