Privacy v. Publicity
User data is valuable. In fact, it is the most valuable resource in the world. Whether it was for searches (Google) or for talking with friends (Facebook), those companies grew according to the amount of user data they collected. Knowing what people think is valuable, what is more private is even more valuable, and when you give people free tools to write that stuff down, and they do, you can make Big Business. When you give people free tools to say that stuff you can make even Bigger Business. The reason is because voice contains more personal, detailed, and valuable data than text. Thus Siri, Cortana, Facebook M, Google Home, Alexa, &co are the mouth, ears, eyes, and sensors of a new surveillance. And surveillance is great if you’re hospitalized or happen to be a parent of a worrisome kid.
What’s with all the fuss over privacy? People say, “I’ve got nothing to hide.” But that just isn’t true (ask them why they’re wearing pants or, better yet, dare them to write down all their passwords and user ids and hand it to you). Everyone has something to hide, otherwise you’re just not a real person. And it turns out that the data policies of hospitals, healthcare centers, nursing homes and many schools are aligned with those of dictator-states, fascist regimes and overly-aggressive marketers.
Privacy is a confusing thing. I’ve had to think of it in simple terms, like an onion ring, in which it’s not binary, but more like a zone. Consider your home. Maybe you have a front yard, or space in front of the apartment building where people kind of walk through and you may never see them. It’s public. Then, entering your home, there’s likely an entry-space that will be occupied by anyone that visits you. Then you have, say your bedroom, now closer to the center of your privacy zone, where very few will visit, and finally your bed is the most private and that is rarely trafficked by others. Chances are you keep something very valuable near your bed. It’s the most private.
Our privacy isn’t a thing that is binary, nor even planar, but volumetric:

Some say that privacy is dead, but I don’t think that’s the case. I just think it’s getting invaded by companies that are profiting from publicizing our private data. Google and Facebook are paid to publicize what we write or say. If we look at it through the eyes of Google it’s a good match, in which the user’s private search is mapped to the marketer’s public advertisement.
Apple is handling this problem via something called “differential privacy” in which they can obscure enough data to not identify an individual, but still identify what people are thinking, writing, and saying. So there’s alternatives in which machine learning can be assisted by privatized data, and that means that there’s solutions we can consider that do a better job than Facebook or Google (and even if you trust Google they can still be hacked and that data you’ve put up will never come down or be forgotten so you never know what will happen to something as permanent and liquid as what you’ve sent to Google 20 years from now).
We need these new models. We need new models both for revenues and for users, that balance publicity and privacy. We need models that allow us to personally set data management policies, allow us to benefit from our own personal data (just as any writer, photographer, or musician would) and models that, in a kind of upstream licensing-rights model, allow us to be remunerated for what privacy we publicize.
Let’s keep thinking on it.
Especially when designing systems as surveillant as bots.
