Oct 17, 2013 · 2 minutes

It's becoming easier and easier to trust our entire lives to technology. The mind-boggling amount of data we produce every day runs through complex algorithms that try to guess what we might want to hear, watch, eat, or do. So many services know what we like well enough that it's easier to let them work their secret magic than to seek new experiences on our own.

Few companies know this better than Stitcher, an Internet radio service that helps its users find new talk shows to which they might like to listen while they work, drive, or walk. The company is today announcing that its mobile applications have been downloaded 12 million times and that the number of shows available through its service has doubled to 20,000 in just one year.

Now that it's reached such scale it plans to improve its personalization features in an effort to make sure that its users never encounter a show -- or an ad -- in which they aren't interested.

"We already know that you're interested in Obama, or you're interested in the government shutdown based on your listening habits," says Stitcher CEO Noah Shanok, "And we're able to deliver more and more content based on that knowledge."

The trouble is balancing those personalization features with Stitcher's attempts to help its users connect with the news. Listening to a certain genre of music to the exclusion of other genres is harmless. Getting news from just one source can be dangerous, not least because of the political partisanship that colors many publishers' reports. As the New York Times' media columnist David Carr put it last week: "It's not just political districts. Our news is gerrymandered too."

As we trust more of our lives to these inscrutable algorithms, it becomes difficult to leave the political echo chamber. The services that predetermine our political leanings don't often encourage us to seek an alternative viewpoint or coverage of an event that may not reflect what we already believe. In other words, it's hard for an algorithm, which is based entirely on past behavior, to encourage a user to broaden his or her horizons.

Stitcher tries to ameliorate this problem by collecting the news and presenting it on its homepage without identifying the source. Users don't know who published the report until they've already decided to play it.

"We run the risk of NPR listeners getting Fox content -- or vice versa -- and that's good," Shanok says. "Some subset of people like that. They want to broaden their comfort zones. Some want to stay within their comfort zone or echo chamber, and we want to be able to provide both."

Of course, that requires us to take some control back from the algorithms. As Stitcher continues to improve its personalization features and expand to other platforms -- such as automobiles, one of Stitcher's most important markets -- that make it harder to fiddle with and find new content that will become harder and harder.

Welcome to your new echo chamber.