A smart speaker has to be listening for its "wake word" to be useful, but there are privacy protections to ensure they don't hear everything you ever say. There are also additional steps you can take to boost privacy levels if you have concerns.
If smart speakers didn't already exist, they would sound like a pretty strange concept. "Hello, we're a giant corporation, and we'd like to put microphones into your home that will listen to you 24/7."
By definition, those microphones are always active. For an Amazon Echo speaker to hear you when you say "Alexa," for example, that microphone has to be switched on and the speaker's computer has to be listening out for that word.
That sometimes leads people to think that a smart speaker is listening to everything they ever say, which isn't entirely true.
With all three of the big ecosystems, they are essentially running the same speech-recognition loop:
Did someone just say my wake word? "Hey, Siri" or just "Siri" "Alexa" or "Ziggy" or "Echo" or "Amazon" or "Computer" "Hey, Google" If no, then forget what was said If yes, listen for a question or commandSo if you're having a conversation, and don't use your smart speaker's wake-word, then it won't process the content of anything you say. If you accidentally say the wake-word, or you say something the speaker thinks is the wake-word, you'll know because it will respond to you.
All smart speakers also have visual indicators to tell you they are processing your speech. These vary between devices, but in all cases, the speaker will use color lighting to indicate that it has been triggered and is now responding.
Apple has the strongest privacy protections of all smart speaker systems. The first of these is on-device speech recognition, where the device itself works out what you have requested. Anything that can be handled locally, is.
That includes smart home commands and setting timers, alarms, and reminders. The actual speech recognition part is also handled on-device for other simple requests like controlling music and asking for news and weather, though of course, the actual data has to be obtained from the internet.
Second, where a request has to be handled by a Siri server, the request is encrypted before it leaves your device. Even if it were somehow possible for a third party to intercept the transmission, it would just be gibberish. The same is true of the response sent back to your device.
Third, the server doesn't know who you are. Instead of sending your Apple ID, your device sends a randomly-created code. There is no way for Apple or anyone else to relate that code to any individual or household.
Apple used to automatically create transcripts of requests, and send a random sample of the recordings and transcripts to human reviewers for quality checking. However, the company ceased doing this by default after it proved controversial, so this no longer happens unless you opt-in.
Amazon's approach to privacy is rather different. Amazon servers do know your identity, and by default the company stores your voice recordings forever. Essentially, Amazon relies on giving users the ability to take control, and these options are discussed below.
Google aims to have the most connected assistant, so by default it knows pretty much everything about you across all the Google services you use. As with Amazon, you are able to take control, though this is much harder with Google as there are so many different services, each with its own privacy policies and controls.
Essentially, if you're the sort of person who wants to set limits to how much your smart speaker assistant knows about you, you probably don't want to choose Google devices.
For Siri, voice recordings are often not sent to Apple servers in the first place. Where they are, your speech is encrypted and anonymized. Recordings are not retained by Apple unless you opt-in.
Amazon keeps your recordings forever, unless you set a different default period or delete them manually.
Google does not store your voice recordings by default.
All three systems offer a number of ways you can boost the privacy of your smart speaker. You can, for example, tell your speaker to cease listening until further notice or delete any stored history.
To cease listening: "Hey Siri, stop listening." You can still make requests by tapping the top of your HomePod, but if you want to re-enable listening, you need to do so through the Home app on your iPhone, iPad, or Mac. Tap the three-dot menu top-right, then Home Settings > Listen for Siri or Hey Siri. You can then toggle listening on and off for each HomePod in your home.
Turning off location awareness is also possible through the same Home Settings menu. Keep scrolling down to Location Services and toggle this off or on. Note that this will apply to all HomePods in your home.
Deleting your Siri history is done on your iPhone, through the Settings app. Scroll down to Siri or Apple Intelligence & Siri (depending on iOS version and iPhone model) then Siri & Dictation History > Delete Siri & Dictation History.
To cease listening: Press the mute button on the speaker. If there is a camera in the device, you can disable this with the camera off button and slider.
To change how long voice recordings are kept: Open the Alexa app on your phone, then Settings > Alexa Privacy > Manage Your Alexa Data. Scroll down to Voice Recordings > Choose How Long to Save Recordings. Deleting your voice recording history is also done here by selecting Don't Save Recordings. This will delete existing ones, and disable saving of future ones.
There are many more privacy options in the Alexa app.
To cease listening: Press the mute button on the speaker. To change how long voice recordings are kept, access your Google Assistant activity then Auto-delete > Auto-delete Activity Older Than and select a time period.
Taking full control over what personal data Google smart speakers use, however, is a nightmare. You'll need to follow each of the individual privacy links in this document, and then follow the instructions in each of those.
Ultimately, only you can decide if you're comfortable with this. Smart speakers are by far the most convenient way to control smart homes, and offer a convenient way to set alarms, create reminders, get weather, and very much more.
Personally, I wouldn't be without mine but I am almost all-in on HomePods, which are the most privacy-focused option. I have one Amazon Echo Dot I use for shopping. I don't have any Google speakers in my home.
There are plenty of ways a smart speaker can help out in your home and boost your productivity, and privacy isn't the only consideration when choosing one—but it may be the most important factor for you.
When you subscribe to the blog, we will send you an e-mail when there are new updates on the site so you wouldn't miss them.
Comments