Wednesday, January 28, 2009

Geotagging is cool…but…

I love to play with new technology, exploring the capabilities and figuring out how it fits into my work and life.  The Eye-Fi memory card is a good example, an SD card for digital cameras with built-in WiFi.  It can upload pictures automatically to phones, computers, or internet albums, and can even imprint location-based information onto pictures based on triangulation against nearby base stations.  The iPhone has similar capability, and it’s spreading to point-and-shoot cameras.

I like the idea of having my pictures automatically tagged with the location of the shot.  I played with various mapping utilities that accompany Flickr, but it’s a lot of effort to position each photo on a map after spending time adjusting color and brightness, tagging them with titles and face identifiers, and organizing them into narrative albums.

More broadly, I like location-aware services.  My phone knows where I can find a shop or restaurant nearby;  TomTom keeps me on track, on time, and ahead of (most of) the speed cameras.  I might be interested in knowing what friends are nearby when I want a social evening at the pub.

All of these are great, but they all involve queries that I make based on geolocation.  When others can find me based on the same technology, I get concerned.  I don’t want advertising from stores aware that I’m walking past.  When Corporate Security assured me that they could follow me in real-time via my cell phone, I almost stopped carrying it.  I’m not sure that I want TomTom querying my navigation system to build up traffic congestion maps.

Matthew Honan wrote another cautionary tale in this month’s Wired Magazine.  He saw a woman taking a picture in a park in San Francisco with her iPhone and guessed that she might upload it to Flickr or Picasa.  Querying based on time and location, he found her pictures, and, based on her identity, could plot out a map of her geotagged uploads.  A characteristic cluster yielded pictures of her apartment, a short jump to address, identity, trouble.

At work, we struggle to find the right balance between out-of-sight convenience and breach-of-privacy intrusion during remote medical monitoring.  I suspect that consumer applications are taking more of an opt-in approach without users realizing the full consequences.  We’ll each have to decide whether we want our complete public trajectory to be publicly accessible; my kids say ‘Why not?’.  I’m not so sure.


Patti said...

But my online photo album is password protected, so that means I'm safe. Right? Right? Oy, gewalt ist mir. I like to think the online info about me is minimal, but then a random, marginal person IRL found my blog from my email address and he was more than a little creepy. Don't ask how he got my email address, at least he didn't figure out where I live. (sitting here shuttering about the possibilities)

Dave Hampton said...

Absolutely: the password is more than enough. Unless your camera(phone) has the feature and does it auto-magically, there's no problem. Even if it does, you can turn it off or generalize the location to a city or state.

'Not good that someone connected the dots; I hope that it doesn't become anything serious.

These convenience features are always two-faced: Facebook status is a wonderful way of touching with friends, but announcing when you're not home to the world is bad. Pictures of lunch in the park is good, but not if it announces why you skipped a meeting.

I always try to know my audience...I use the privacy settings, and keep my travel posts retrospective on the blog. It's not a worry: it's just something to be aware of and manage. Still, the Wired article caught me by surprise.