If every tree falling in every forest might soon be heard by an internet-connected microphone, what hope is there for our privacy?
Already when you’re sitting in a room with an iPhone, an Apple Watch and a smart assistant like Amazon Echo or Google Home, you’re surrounded by a dozen microphones. (Newer iPhones have four and the Echo has seven, while the smartwatch has just one, for now.)
Add in the latest smart wireless headphones—Apple’s expected next-generation AirPods or competing ones from Bose or Shure—along with talking microwave ovens and TVs from Samsung, LG and others, and anyone at home or in an open-plan office could soon be within earshot of hundreds of microphones. Most of them will be listening for a wake word like “Alexa,” “Hey Siri,” or “OK Google,” just as our phones and smart assistants do now.
The roadmaps of tech giants and startups alike show how sound is poised to become the first ubiquitous connection between users and the artificial-intelligence hive mind the internet is becoming.
Driving this change are massive volumes of components, originally designed for smartphones and other mobile devices, that have dropped in price and grown in functionality over the past decade.
For a hundred years, microphones consisted of a relatively large membrane whose vibrations were converted to electrical impulses. But starting in the 1980s, engineers worked out ways to make microphones tiny, bordering on microscopic. Most still have a pocket of air trapped behind a vibrating element, but now they can be carved out of silicon, just like the microchips to which they’re attached. Smartphones, smart speakers and any other gadget that listens for your voice all use these kinds of microphones.
, based in Itasca, Ill., has more than 50% of this market. Selling to all major manufacturers of mobile devices, the company has shipped 12 billion of them over the past decade, says a company spokesman. (A long list of other microphone suppliers, including
AAC Technologies and STMicroelectronics, constitute the remainder of that market.)
One ongoing challenge for microphones has been physics: The smaller microphones get, the more of them you need to capture a sound, and the more processing of that sound is required.
Startups such as Boston-based Vesper Technologies, Inc.—which has received money from Baidu, Bose and Amazon’s Alexa Fund—are meeting the challenge with even tinier, yet more capable designs built around minuscule flaps of silicon that generate electric current when bent by sound waves. Vesper claims this gives their microphone unique capabilities, like understanding your voice even in windy conditions, and drawing zero power when awaiting a “wake word,” since sound itself generates the power the microphone needs.
The total cost to equip a gadget with an array of these tiny microphones and the electronics to interpret simple commands is approaching $10 or less, says Matt Crowley, Vesper’s chief executive. Individual microphones now cost between 20 cents and 60 cents, says
an analyst and marketing chief at Applied Materials Inc., which supplies manufacturing equipment to makers of microchips.
We’re moving toward a world in which everything with a plug or battery can respond to a voice command.
Apple’s next AirPods could have many of the capabilities that Vesper claims its microphones will enable, such as built-in noise cancellation. (In the past, Apple has used several suppliers for its microphones.) Meanwhile, the CEO of Samsung’s consumer-electronics division recently told The Wall Street Journal that by 2020 his company plans to equip every single device it sells—from TVs to refrigerators—with microphones.
It could be unnerving to be surrounded by listening devices, but the paradox is that as the technology develops, so does our ability to free these gadgets from having to connect to the internet.
Consider the voice-controlled trash can from Simplehuman. Say “Open can” and it opens—and then closes on its own once the user walks away. That’s it.
While it’s easy to make fun of a high-tech trash can, especially one that costs $200, this one tackles one of the biggest concerns that comes with smart assistants: the fact that they record what we tell them and send it back to their parent companies.
Simplehuman’s trash can doesn’t do this, says
the company’s director of electronics engineering. That’s because the latest microphones and their attached microprocessors process human speech in the gadget itself, without connecting to the cloud.
At first, self-contained processing in gadgets will be limited to simple commands and wake words, like telling a device to turn on or setting a timer. In time, these commands will become more complex.
One justification for adding voice control to everything we use is that it could ultimately be an easier and more elegant interface than the morass of buttons and menus we face today. Just think how hard it is to work a friend’s microwave, convection oven or thermostat, and imagine instead just telling it what you’d like it to do.
A future full of always-listening devices will have its own complications, of course. One challenge will be the necessity of all of us going through our days constantly muttering to ourselves, or projecting our voices at tin-eared appliances.
As anyone who lives with multiple virtual assistants can attest, it is tricky to talk to one without inadvertently involving the whole crowd. Simplehuman admits that its trash can sometimes spontaneously opens in response to background noise. Even Amazon’s Echo wakes up when it isn’t needed, and can misinterpret family chatter as a directive to fire off seemingly random messages.
When we leave instructions for dog sitters or house guests, they include notes about the quirks of the interfaces to our appliances, gadgets and heating and cooling systems. In the future, they might consist solely of a list of names for all those devices, customized to reflect our tastes.
It might not be long before you find yourself saying something like, “David Bowie, preheat the oven to 350 degrees. Frank Zappa, wash the dishes.”
Write to Christopher Mims at firstname.lastname@example.org