Thursday, September 09, 2004
I don't recall if I have mentioned this before (and I'm too lazy to go through all the entries to check), but here are two of my predictions for the near future:
- Within 5 years, technology will be packaged in a way that allows consumers to record audio for every waking moment of their lives
- Within 10 years, the same will be true for video
I'll try to go into depth on each of these predictions, and related outcomes in a future entry, but for the moment, let's assume that constant video recording becomes a fact at some point in the future. I'm also a believer that we will do far more with this video than just record it (for example, I think the ability to search and playback this video on the fly will have wide ranging social impacts); we will have software that analyzes this video in real time for a number of different purposes (again, more in a later entry), one of which will be facial recognition. Another is likely to be optical character recognition (couple OCR with GPS,and my quick entry about the gas prices information swarm, and you get automated gas price reporting without requiring station owners to use a central system).
Facial recognition will likely come about as a means to automatically look up information that we have stored about a person we are looking at. Remembering their name would be a fabulous application for me, and remembering such things as their birthday, or kids names and ages and the like is probably going to make the application worthwhile enough to at least be investigated.
So it occurred to me a couple of weeks ago as I was walking down the street that if I had a camera headset on that had OCR, it could scan license plates, which might be handy for spotting a friend's car. Of course the much more obvious use is in law enforcement. Given limited budgets and staff, already a number of police departments are reaching out to the community using new-ish technology to cover more area more economically. I heard a story on NPR last week about a community in Florida (I think) that has recruited a number of people in the community to be part of their cell phone network. Essentially whenever one of these people (whom I presume do not get paid) sees something out of the ordinary, they call in to a special number to report it. The police department can then dispatch enforcement officers to those things that might warrant it, and track bits of information that may not mean anything singularly, but may form a pattern over time.
So if I have a camera that can perform OCR in real time, then it seems like the police might want to tap into that information to try and locate a stolen car, for example. If this OCR data is recorded, then the police might want to tap into your history to find out if you ever saw a car that was used in a crime to try and discern whose car it was as part of their evidence chain.
Furthermore, this scenario extends even further when you throw in facial recognition. Police would want to identify where a suspect was last seen, if a missing child had been spotted, or the current whereabouts of persons of interest.
My guess is that the desire to tap this information network will create huge pressures to legislate police access to your data in real time. Rather than calling to law enforcement every time a plate is read or a face is matched, a solution might be to have a central clearinghouse and a voluntary checks against it from your gear. This would allow enough "people on the streets" to help out law enforcement while not being overly Orwellian. Preferably, your gear would download lists of plates and facial markers (probably in encrypted form so you couldn't look at the list) allowing your system to check for matches locally without calling back to a central location and increasing privacy issues.
Ideally, such an increase in surveillance would come with increased transparency to the organizations that use the system, to help fend off abuse. But I'm guessing that that won't happen in any meaningful way in association with this new information gathering technique. Optimistically, that might be because the current overhaul of the intelligence agency might foster such transparency before information networks like this come online in a meaningful way. More realistically, however, I think that without a major public abuse to trigger public interest, not enough people will really care (note that this is different from not enough people having been exposed to the issues) enough to pressure for change.
And this is just one of the interesting social struggles we are likely to have as computing power, portability, and ubiquity really hit their strides. More later ;)