Evernote, Humans, and Machines

The recent Evernote privacy policy brought about considerable response from the nerds.

Here’s the relevant section many are concerned with (highlighting is mine):

The latest update to the Privacy Policy allows some Evernote employees to exercise oversight of machine learning technologies applied to account content, subject to the limits described below, for the purposes of developing and improving the Evernote service.

Now I’ve highlighted two points in the policy. The part most people are concerned about is the bit where an Evernote employee reviewing customer data. That’s certainly unsavory if you are cataloging your entire life. But it’s the second part that I’d be concerned with if I were still an Evernote customer: “…machine learning technologies applied to account content…”

It’s somewhat strange to me that we would all be concerned about one unknown human actor reviewing our information from thousands of miles away. I’m far more concerned about a learning algorithm connecting every bit of information I capture. I’m sure the machine learning will provide some very cool features but I also think that a good algorithm could automatically mine and identify aspects of my life I don’t want used outside of my control.

Evernote has updated their policy to remove the human oversight. Two steps back, one step forward.