In 2014, Amazon debuted a simple but industry-changing product: the smart speaker. Technically the Amazon Echo was just a microphone attached to the internet that you installed in your home. But it let users ask a digital assistant, Alexa, thousands of questions and commands, and it was a hit. Before long, Google and Apple followed with their own smart speakers, and today, a device that began as a curiosity has become commonplace: one in five US households now owns a smart speaker.
Smart speakers offer convenience; much of their popularity can simply be chalked up to that. But tech companies are also clearly pushing the technology onto consumers hard, sometimes selling smart speakers at rock-bottom prices, and building the “listening” technology that drives them into all sorts of other products, from headphones to doorbells. And a big reason for that is all the data that they produce. Just like our web searches, online purchases, and social networks, every command you give to a smart speaker is a new piece of data that tech companies own. Most likely, your voice recordings are already being used for improving those companies’ listening algorithms and ad targeting, but there’s very little transparency and no way to know exactly how they use human voice data. All we really know is that these devices have enabled their manufacturers to collect gigantic troves of voice recordings — and that opting out of it isn’t always easy. Open Sourced is a year-long reporting project from Recode by Vox that goes deep into the closed ecosystems of data, privacy, algorithms, and artificial intelligence. Learn more at http://www.vox.com/opensourced This project is made possible by the Omidyar Network. All Open Sourced content is editorially independent and produced by our journalists.