A private function

Features editorAs sensors become cheaper and more ubiquitous, do we need to be more careful about our privacy — or do we need to redefine what we mean by ‘private’?

The latest version of Microsoft’s XBox gaming system is raising eyebrows among its target market. Among its features is an integrated Kinect system — the sensor array which allows users to interact with games by moving or speaking, rather than by operating a controller. But even when the console is switched off, the Kinect unit’s microphone remains on, as part of a voice-activated ‘on-switch’. The microphone is not recording in this state, and neither is it connected to the internet — but many people are worried that an always-on microphone represents an invasion of privacy.

Now, that’s an interesting point. Ask a gamer a few years ago whether they’d like a system they could switch on by talking to it, and they might well have been enthused by the science-fictiony idea of it, without thinking that it meant the system would have to be listening out for the activation phrase all the time. But as sensor systems which detect aspects of our interaction with the environment become more common and more complex, we might have to ask ourselves some difficult questions about privacy — starting with what, precisely, we mean by ‘private’.

Is the new Xbox invading your privacy? Probably not, if the microphone isn’t transmitting in any way. But it is capable of transmitting, and covert listening (or viewing, through the system’s built-in camera) might be possible. Similar concerns have been raised about Google’s wearable Glass augmented reality system, which is similarly equipped with a camera and systems to record video and log it on-line.

This certainly seems like a valid area of concern. But what about other sensor systems? What about smart metering, for example? A system which transmits your energy usage back to the electricity supplier for accurate metering is great for convenience, but it could also be seen as an invasion of privacy — every appliance you use has a characteristic signature energy usage, so it might be possible for somebody to look at the energy usage and tell when the television is switched on. Or even when the fridge door opens.

Digital healthcare, to some, also represents an invasion of privacy, as it can use sensors which transmit or record measurements such as temperature or blood pressure, or readings which the patient has taken such as blood sugar levels. But what if the transmission only occurs when an algorithm detects an abnormality in the readings? And as these systems are used to maintain or improve patient health, does it even count as invasive?

Systems I’ve mentioned here are ones which are installed in the home; if you’re inviting the sensors in, then you could be said to have consented to their data-gathering. But sensors are becoming increasingly ubiquitous in public spaces as well, and that’s another matter. Clearly there is an issue of privacy here, and it’s one which needs to be discussed.

What information are we happy to have gathered and recorded? What uses are we happy for this information to be put to? If we gain some advantage from the data transmission — for example, recording of our position by GPS to provide us with area-specific information such as road conditions, weather reports and local amenities — does that justify the use of sensor networks, as Chris Parker of Ordnance Survey explained in this interview? How much information is too much? How personal is too personal? All of these are difficult questions, and I certainly can’t claim to have any answers. Not being a gamer myself, I’m not likely to have an Xbox in my home — but I can’t say I’m too keen on the idea of voice-activated on-switches in internet-capable devices.