30.11.2011 - Unintended Consequences of Visual Inputs

I was originally going to write about how QR codes stink but I think there's enough people out there that already share that sentiment. Then an episode of Spark got me thinking about visual inputs in general and how QR codes are yet another type of visual input along with bar codes, facial recognition, and landmark recognition.
This trend in creating computers to take on a more "human" quality isn't new. Engineers and developers have been slowly tyring out different ways to provide computers with input that go beyond the traditional "push-button" method. These include speech recognition (widely accepted), bio-metric recognition (fairly new), and bio-chemical recognition (coming soon?).
Computer inputs such as these are complicated and recognizing them is even more complex. Because of this there can be some unintended consequences of using this type of technology:
1. The new Nexus Galaxy allows users to use facial recognition to unlock their phones but doesn't differentiate between your actual face and simply a picture of your face:

2. Not only can QR codes contain ASCII data, it can also contain binary data. This carries the potential for users to inadvertently scan a QR code that contains malicious code. On a simpler level, a QR code can direct you to a website where malware is fed to your phone. Such was the case as demonstrated by Augusto Pereyra in Russia.
3. With traditional manual inputs you decide when and where you want to provide the computer with input. Now, especially in "smart homes" that have their own system for hearing and seeing users, the computer is always listening and watching users for a recognizable command. Real Big Brother type stuff that seems—to me at least—a bit creepy.
There's no doubt that these alternative inputs need further development. I can't help but gain an appreciation for the 4 billion years of evolution that allowed our brains to seamlessly process our surroundings. Facial recognition is one thing, but what about recognizing facial emotions? Or recognizing someone's face after it has aged 20 years? Or even recognizing a sibling based on facial features? Perhaps on a more esoteric level, how do we create computers to recognize the individual as a whole rather than the parts and features that make the individual? This type of Gestalt approach to computing and technology forces us to ask many difficult questions. What makes a person a person? And more practically, can it be taught to a computer?
Jeff