Is the thermostat creepy?

Posted by on Aug 21, 2013 in Analytics

In this age of the internet of things, it seems every new device made today wants to connect to the internet and share data about our surrounding world. Which means that we will (and in many cases are) have more and more “recommendations” delivered to us and some of them are disturbing or even “creepy”.

This happened to me the other day when my android phone told me the best route and time to get to a store I hadn’t realized I wanted to shop at.  On further consideration, the inference made sense; I had been searching for some products that Ethernet store in question carried, but I never considered that store, in fact I didn’t really want the product right now, but was just checking it out.

That creepy feeling started me thinking about why.

It has always bothered me when a television show I was watching had commercials that ALL were about products in which I had absolutely no interest.  Was it me that was out of step, or the advertisers? When you market to the middle or focus only broad demographic targeting, there will be an inevitable mis-match of product and individual. But online, when I would search for a restaurant and saw an ad that showed exactly what I wanted, I was pleased.

So what causes crossing the line from helpful to creepy?  I have long used the example of the thermostat, and its audacity to monitor my house and adjust the furnace without asking. It was doing what I wanted and I was glad.  The newer thermostats allow me to tell it what temperature I want at different times of the day or even weekends.  nestNow we have the “NEST” that pays attention to how I change it and generate its own rules to make me happy, still all good.

But, what if there where sensors that measured who was in the house and what it was like outside and even the weather forecast, had nano sensors embedded in the occupants, and could tell when they were working out or knew from their calendar that there was a party tonight and adjusted the temperature accordingly. Creepy? Yes and no. You may be glad it is doing the right thing, but slightly uncomfortable about how much it knows about your habits!

We are happy when our abs brakes prevent an accident, but may get upset when we get a discount coupon for a baby gift for our single daughter who herself doesn’t know yet that she is pregnant.  Additionally, we have bought the thermostat and the ABS and given it permission to monitor us and the situation and make good decisions.  The challenge to the decision scientist is to provide valuable information without disturbing the customer or violating his desires and permissions; this will likely require direct feedback.

Finally, please note that I’m not talking about the “Singularity” – the point in which artificial intelligence matches and surpasses human intellect, and then machines take halover the earth and wage war with the human race, etc, etc. I’ve seen that movie (more than a few times) and, frankly, it’s just too difficult to predict what that type of impact future machine intelligence would have on whatever state the world is in when we reach that point. I’d prefer to deal in the now and in the near future. The systems I’ve mentioned all exist today, and every major electronics company seems to be working furiously to make tomorrow’s gadgets and tools smarter, more connected, and more personalized. The fact that some of these may elicit a creepy feeling from us is ok. It just means we’re human and need to keep rebalancing the line between automation thats helpful and that which thinks for us.