[Cross-posted from blog.davemdavis.net]
Over the past two posts I started to explore the concept of ubiquitous computing. I started out by defining ubiquitous computing, and then followed that up with a post that talks about how developers can contribute to the ubiquitous utopia. No conversation about ubiquitous computing would be complete without talking about privacy. Privacy is a major concern these days, especially with all the controversy surrounding NSA data collection programs.
Recently, at work, we discussed the article AI is so over: This is artificial empathy. This article talks about a system that can determine a person’s personality based upon a few seconds of speech. This system is used by call centers to determine the mood of the caller and then to direct the caller to a representative that can best handle him or her. Let’s think about that for a minute. From a few seconds of voice analysis, this system has searched its vast amount of data, analyzed the speech, and accurately determined a caller’s personality. In order to continually learn, the system has to collect and analyze a large amount of voice samples.
Who owns this data? It is your voice they are collecting? Can you get the data back? “This call may be monitored for quality assurance.” Currently this system is used for customer services. It may even provide the benefit of a smooth customer service experience. What happens when they start using this system in other areas? Can an organization refuse a job because you come off as being mean? Can/should an adoption agency deny an applicant because they don’t have the right personality as determined by a few seconds of speech?
Let’s say that you agree there is some benefit of relinquishing some of your privacy for the betterment of the system. How much privacy are you willing to part with? In order for ubiquitous computing to work, we must be willing to sacrifice some privacy. These computing systems rely on context to perform their task while remaining in the background. The more information you provide, the better the chance the system get the right context, and the richer the experience can be.
Ubiquitous in the Home
The previous example may be a little too far fetched for people to grasp. Who actually calls customer support, these days, anyway? A lot of ubiquitous computing examples can be found in home automation. One of the latest fads in home automation is the Nest Thermostat. This thermostat learns. It gathers usage data to determine patterns of how you heat or cool your home. If you set the thermostat to a specific temperature every morning when you wake up, it learns that time. It also knows how long it will take to heat/cool your house to that temperature. After a while (a week according to the web site) it will start to turn itself on so that your house will be at the desired temperature around the time you wake up.
The Nest has sensors that can determine when you are home or not and then set the temperature of your house appropriately. If you have multiple zones in your house and you use multiple Nest devices, they can communicate with each other, connecting through your Wi-Fi network. There are even mobile applications that allow you to control your thermostats remotely. If you are coming home early you can kick the heat on to warm up the place. So what is the benefit to fully embracing the Nest system? Reduced energy bills. You no longer have to remember to turn down the heat. The Nest does that by learning your habits and automatically adjusting to an appropriate temperature. Over time you should see a reduction in your bill as the system is only on when needed. So is the slight invasion of privacy worth saving some money in your bill?
Should I Care
On the surface it seems harmless that Nest is collecting all this data but let’s take a deeper look. The system knows when you are home or away and keeps a record of that. Do you think that burglars would like that information? Oh did I forget to mention that Google just purchased Nest for $3.2 billion dollars. That sales price is 10x the sales figures for Nest. Why would Google pay that much for Nest? Could it be because it wants the data that is collected? This is just another touch point that Google has on your life. Add this data with other data they collect and you have a pretty good picture into your life. A picture that advertisers would kill for.
So What Can I Do
As a developer, you should be aware of what personal information your system requires in order to gain a usable context. Understand that some users will be very apprehensive to give up their privacy to your system. Provide the user with some assurance on how you plan to use their data. If you plan on changing your data collection policies, provide enough notice so that the user can make an informed decision. Finally provide a mechanism for the user to take back their data should they decide to stop using your service/system/device.
As a consumer be conscious of the information you provide. Keep in mind that the information you provide one company may be aggregated with information you provide to a different company. Weigh the consequences of relinquishing your privacy vs. the benefit provided by the system. Remember these interconnected systems require your personal information in order to provide an accurate context, so that they can deliver an optimal experience.
This post covered two examples of where your privacy is affected by ubiquitous computing. There are many more examples, just look around. In a future post I will look into what it means for a system to have context and how context drives the notion of ubiquitous computing.