Avoid creepiness

A couple weeks ago, the New York Times had a pretty decent article about the internet of things (IoT), The Internet Gets Physical. The focus was on the more industrial side of things as opposed to Arduinos and the maker culture.

And in between a number of pretty awesome-looking projects like the UN’s Global Pulse, I noticed this (highlights are mine):

Researchers at General Electric, the nation’s largest industrial company, are working on such applications and others. One is a smart hospital room, equipped with three small cameras, mounted inconspicuously on the ceiling. With software for analysis, the room can monitor movements by doctors and nurses in and out of the room, alerting them if they have forgotten to wash their hands before and after touching patients — lapses that contribute significantly to hospital-acquired infections. Computer vision software can analyze facial expressions for signs of severe pain, the onset of delirium or other hints of distress, and send an electronic alert to a nearby nurse.

So now there we have an interesting case. A scenario that both makes rational sense and is at the same time absolutely unthinkable. Yes, in this scenario we would use technology to reduce risk for patients. But it would come at the cost of the most severe privacy violations, particularly in a situation as vulnerable as a hospital stay.

The fact that the cameras are mounted inconspicuously on the ceiling only reaffirms the impression that they shouldn’t be there in the first place. (I also doubt that facial expression would even work reliably when, for example, you’re attached to a respirator, but that isn’t even the point.)

For IoT applications to get more widely accepted, they should be more human, humane, and human-friendly. The Nest thermostat is a good example. It’s helpful, looks good and is decidedly non-creepy.

Being non-creepy is good.


what’s interesting is that all of the goals can be achieved by different means than cameras. want doctor & nurse tracking? rf triangulation. pain level measurement? there’s galvanic skin response sensors. so you have to think about why they’d put up cameras. (sensors are btw way cheaper than the software to analyse the use cases laid out here.)

what I find more disturbing though is the apparent focus on all things surveillance when it comes to IoT. Monitor patients, your home, your car. monitor, monitor, monitor.

we desperately need a better narrative.

yeah, sure. I’m not arguing that surveillance and monitoring applications don’t make sense in a lot of cases. I’m arguing against monitoring and surveillance as the MAIN SELLING POINT of sensors and IoT roll out. This creates a narrative which emphasizes control whereas it should emphasize empowerment.

Leave a Reply