We’re all aware that privacy is essentially nonexistent on today’s web. Using a VPN and switching to Firefox may help, but it’s a sad fact of life that advertisers follow us from site to site, creating personas of who we are and offering up just the right ad in response.
Increasingly, AI researchers are pushing the boundaries of how they can identify individuals in real life, too. Soon, architecture itself could help organizations, governments, and companies track people as they move through the world–and not just through cameras. From your snorts to your gait, software may soon be able to decipher your identity in all sorts of crazy ways. The good news is that we’re all unique snowflakes after all. The bad news is that in the hands of neural networks, those snowflakes are very, very easy to recognize.
The way you walk
People have at least 24 different factors that can be measured while they walk. A new system called SfootBD, published this spring in the journal IEEE Transactions on Pattern Analysis and Machine Intelligence, can track them all. That makes the system capable of creating a unique signature of your gait, allowing it to identify who you are based on your footsteps. The hardware system itself requires pressure sensors in the floor and a high-resolution camera, but these items can be installed relatively covertly. With a 0.7% rate of error, it’s not perfect, but it’s shockingly close. Oh, and Wi-Fi signals can identify your walking, too.
The way you type
The way you talk
Platforms like Google Home can already distinguish Mom from Dad in a home. But cutting-edge voice research can do a lot more than that. Chase and Wells Fargo are both using “voiceprinting” to identify frequent fraudulent callers. One researcher recently likened voice to “your fingerprints or DNA.” But even if it’s not a perfect portrait, voice can still help identify someone’s height, weight, and demographic–even their psychological state
The way you breathe between words
Gasp. By recording the inhalation and air turbulence caused by breathing in between the words we speak, Carnegie Mellon researchers have learned how to identify individual people with 91.3% accuracy. It sounds crazy, until you realize that everyone’s oral passageway varies, like a uniquely shaped instrument. Of course, your breath can give you away in other ways, too. While it doesn’t rely upon AI, scientists can measure the chemistry in your exhalation–a breathprint–and identify you and the unique mix of microbes inside of you.
Combined, all of these techniques are enough to make you wonder if we’ll ever be able to guarantee your privacy, or anonymity, again. Without more regulations in place, a GDPR for every building in the entire world, perhaps, it certainly doesn’t seem very likely.