by Kevin Kelly
Still, the picture is not big enough. We—the internet of people—will track ourselves, much of our lives. But the internet of things is much bigger, and billions of things will track themselves too. In the coming decades nearly every object that is manufactured will contain a small sliver of silicon that is connected to the internet. One consequence of this wide connection is that it will become feasible to track how each thing is used with great precision. For example, every car manufactured since 2006 contains a tiny OBD chip mounted under the dashboard. This chip records how your car is used. It tracks miles driven, at what speed, times of sudden braking, speed of turns, and gas mileage. This data was originally designed to help repair the car. Some insurance companies, such as Progressive, will lower your auto insurance rates if you give them access to your OBD driving log. Safer drivers pay less. The GPS location of cars can also be tracked very accurately, so it would be possible to tax drivers based on which roads they use and how often. These usage charges could be thought of as virtual tolls or automatic taxation.
* * *
• • •
The design of the internet of everything, and the nature of the cloud that it floats in, is to track data. The 34 billion internet-enabled devices we expect to add to the cloud in the next five years are built to stream data. And the cloud is built to keep the data. Anything touching this cloud that is able to be tracked will be tracked.
Recently, with the help of researcher Camille Hartsell, I rounded up all the devices and systems in the U.S. that routinely track us. The key word is “routinely.” I am leaving off this list the nonroutine tracking performed illegally by hackers, criminals, and cyberarmies. I also skip over the capabilities of the governmental agencies to track specific targets when and how they want to. (Governments’ ability to track is proportional to their budgets.) This list, instead, tallies the kind of tracking an average person might encounter on an ordinary day in the United States. Each example has been sourced officially or from a major publication.
Car movements—Every car since 2006 contains a chip that records your speed, braking, turns, mileage, accidents whenever you start your car.
Highway traffic—Cameras on poles and sensors buried in highways record the location of cars by license plates and fast-track badges. Seventy million plates are recorded each month.
Ride-share taxis—Uber, Lyft, and other decentralized rides record your trips.
Long-distance travel—Your travel itinerary for air flights and trains is recorded.
Drone surveillance—Along U.S. borders, Predator drones monitor and record outdoor activities.
Postal mail—The exterior of every piece of paper mail you send or receive is scanned and digitized.
Utilities—Your power and water usage patterns are kept by utilities. (Garbage is not cataloged, yet.)
Cell phone location and call logs—Where, when, and who you call (metadata) is stored for months. Some phone carriers routinely store the contents of calls and messages for days to years.
Civic cameras—Cameras record your activities 24/7 in most city downtowns in the U.S.
Commercial and private spaces—Today 68 percent of public employers, 59 percent of private employers, 98 percent of banks, 64 percent of public schools, and 16 percent of homeowners live or work under cameras.
Smart home—Smart thermostats (like Nest) detect your presence and behavior patterns and transmit these to the cloud. Smart electrical outlets (like Belkin) monitor power consumption and usage times shared to the cloud.
Home surveillance—Installed video cameras document your activity inside and outside the home, stored on cloud servers.
Interactive devices—Your voice commands and messages from phones (Siri, Now, Cortana), consoles (Kinect), smart TVs, and ambient microphones (Amazon Echo) are recorded and processed on the cloud.
Grocery loyalty cards—Supermarkets track which items you purchase and when.
E-retailers—Retailers like Amazon track not only what you purchase, but what you look at and even think about buying.
IRS—Tracks your financial situation all your life.
Credit cards—Of course, every purchase is tracked. Also mined deeply with sophisticated AI for patterns that reveal your personality, ethnicity, idiosyncrasies, politics, and preferences.
E-wallets and e-banks—Aggregators like Mint track your entire financial situation from loans, mortgages, and investments. Wallets like Square and PayPal track all purchases.
Photo face recognition—Facebook and Google can identify (tag) you in pictures taken by others posted on the web. The location of pictures can identify your location history.
Web activities—Web advertising cookies track your movements across the web. More than 80 percent of the top thousand sites employ web cookies that follow you wherever you go on the web. Through agreements with ad networks, even sites you did not visit can get information about your viewing history.
Social media—Can identify family members, friends, and friends of friends. Can identify and track your former employers and your current work mates. And how you spend your free time.
Search browsers—By default Google saves every question you’ve ever asked forever.
Streaming services—What movies (Netflix), music (Spotify), video (YouTube) you consume and when, and what you rate them. This includes cable companies; your watching history is recorded.
Book reading—Public libraries record your borrowings for about a month. Amazon records book purchases forever. Kindle monitors your reading patterns on ebooks—where you are in the book, how long you take to read each page, where you stop.
Fitness trackers—Your physical activity, time of day, sometimes location, often tracked all 24 hours, including when you sleep and when you are awake each day.
It is shockingly easy to imagine what power would accrue to any agency that could integrate all these streams. The fear of Big Brother stems directly from how technically easy it would be to stitch these together. At the moment, however, most of these streams are independent. Their bits are not integrated and correlated. A few strands may be coupled (credit cards and media usage, say), but by and large there is not a massive Big Brother–ish aggregate stream. Because they are slow, governments lag far behind what they could do technically. (Their own security is irresponsibly lax and decades behind the times.) Also, the U.S. government has not unified these streams because a thin wall of hard-won privacy laws holds them back. Few laws hold corporations back from integrating as much data as they can; therefore companies have become the proxy data gatherers for governments. Data about customers is the new gold in business, so one thing is certain: Companies (and indirectly governments) will collect more of it.
The movie Minority Report, based on a short story by Philip K. Dick, featured a not too distant future society that uses surveillance to arrest criminals before they commit a crime. Dick called that intervention “pre-crime” detection. I once thought Dick’s idea of “pre-crime” to be utterly unrealistic. I don’t anymore.
If you look at the above list of routine tracking today, it is not difficult to extrapolate another 50 years. All that was previously unmeasurable is becoming quantified, digitized, and trackable. We’ll keep tracking ourselves, we’ll keep tracking our friends, and our friends will track us. Companies and governments will track us more. Fifty years from now ubiquitous tracking will be the norm.
As I argue in Chapter 5 (Accessing), the internet is the world’s largest, fastest copy machine, and anything that touches it will be copied. The internet wants to make copies. At first this fact is deeply troubling to creators, both individual and corporate, because their stuff will be copied indiscriminately, often for free, when it was once rare and precious. Some people fought, and still fight, very hard against the bias to copy (movie studios and music labels come to mind) and some people chose and choose to work with the bias. Those who embrace th
e internet’s tendency to copy and seek value that can’t be easily copied (through personalization, embodiment, authentication, etc.) tend to prosper, while those who deny, prohibit, and try to thwart the network’s eagerness to copy are left behind to catch up later. Consumers, of course, love the promiscuous copies and feed the machine to claim their benefits.
This bias to copy is technological rather than merely social or cultural. It would be true in a different nation, even in a command economy, even with a different origin story, even on another planet. It is inevitable. But while we can’t stop copying, it does matter greatly what legal and social regimes surround ubiquitous copying. How we handle rewards for innovation, intellectual property rights and responsibilities, ownership of and access to the copies makes a huge difference to society’s prosperity and happiness. Ubiquitous copying is inevitable, but we have significant choices about its character.
Tracking follows a similar inevitable dynamic. Indeed, we can swap the term “tracking” in the preceding paragraphs for “copying” in the following paragraphs to get a sense of its parallels:
The internet is the world’s largest, fastest tracking machine, and anything that touches it that can be tracked will be tracked. What the internet wants is to track everything. We will constantly self-track, track our friends, be tracked by friends, companies, and governments. This is deeply troubling to citizens, and to some extent to companies as well, because tracking was previously seen as rare and expensive. Some people fight hard against the bias to track and some will eventually work with the bias. Those who figure out how to domesticate tracking, to make it civil and productive, will prosper, while those who try only to prohibit and outlaw it will be left behind. Consumers say they don’t want to be tracked, but in fact they keep feeding the machine with their data, because they want to claim their benefits.
This bias to track is technological rather than merely social or cultural. It would be true in a different nation, even in a command economy, even with a different origin story, even on another planet. But while we can’t stop tracking, it does matter greatly what legal and social regimes surround it. Ubiquitous tracking is inevitable but we have significant choices about its character.
* * *
• • •
The fastest-increasing quantity on this planet is the amount of information we are generating. It is (and has been) expanding faster than anything else we can measure over the scale of decades. Information is accumulating faster than the rate we pour concrete (which is booming at a 7 percent increase annually), faster than the increases in the output of smartphones or microchips, faster than any by-product we generate, such as pollution or carbon dioxide.
Two economists at UC Berkeley tallied up the total global production information and calculated that new information is growing at 66 percent per year. This rate hardly seems astronomical compared with the 600 percent increase in iPods shipped in 2005. But that kind of burst is short-lived and not sustainable over decades (iPod production tanked in 2009). The growth of information has been steadily increasing at an insane rate for at least a century. It is no coincidence that 66 percent per year is the same as doubling every 18 months, which is the rate of Moore’s Law. Five years ago humanity stored several hundred exabytes of information. That is the equivalent of each person on the planet having 80 Library of Alexandrias. Today we average 320 libraries each.
There’s another way to visualize this growth: as an information explosion. Every second of every day we globally manufacture 6,000 square meters of information storage material—disks, chips, DVDs, paper, film—which we promptly fill up with data. That rate—6,000 square meters per second—is the approximate velocity of the shock wave radiating from an atomic explosion. Information is expanding at the rate of a nuclear explosion, but unlike a real atomic explosion, which lasts only seconds, this information explosion is perpetual, a nuclear blast lasting many decades.
In our everyday lives we generate far more information that we don’t yet capture and record. Despite the explosion in tracking and storage, most of our day-to-day life is not digitized. This unaccounted-for information is “wild” or “dark” information. Taming this wild information will ensure that the total amount of information we collect will keep doubling for many decades ahead.
An increasing percentage of the information gathered each year is due to the information that we generate about that information. This is called meta-information. Every digital bit we capture encourages us to generate another bit concerning it. When the activity bracelet on my arm captures one step, it immediately adds time stamp data to it; it then creates yet more new data linking it to other step bits, and then generates tons of new data when it is plotted on a graph. Likewise, the musical data captured when a young girl plays her electric guitar on her live video stream becomes a foundation for generating indexing data about that clip, creating bits of data for “likes” or the many complex data packets needed to share that among her friends. The more data we capture, the more data we generate upon it. This metadata is growing even faster than the underlying information and is almost unlimited in its scale.
Metadata is the new wealth because the value of bits increases when they are linked to other bits. The least productive life for a bit is to remain naked and alone. A bit uncopied, unshared, unlinked with other bits will be a short-lived bit. The worst future for a bit is to be parked in some dark isolated data vault. What bits really want is to hang out with other related bits, be replicated widely, and maybe become a metabit, or an action bit in a piece of durable code. If we could personify bits, we’d say:
Bits want to move.
Bits want to be linked to other bits.
Bits want to be reckoned in real time.
Bits want to be duplicated, replicated, copied.
Bits want to be meta.
Of course, this is pure anthropomorphization. Bits don’t have wills. But they do have tendencies. Bits that are related to other bits will tend to be copied more often. Just as selfish genes tend to replicate, bits do too. And just as genes “want” to code for bodies that help them replicate, selfish bits also “want” systems that help them replicate and spread. Bits behave as if they want to reproduce, move, and be shared. If you rely on bits for anything, this is good to know.
Since bits want to duplicate, replicate, and be linked, there’s no stopping the explosion of information and the science fiction levels of tracking. Too many of the benefits we humans covet derive from streams of data. Our central choice now is: What kind of total tracking do we want? Do we want a one-way panopticon, where “they” know about us but we know nothing about them? Or could we construct a mutual, transparent kind of “coveillance” that involves watching the watchers? The first option is hell, the second tractable.
Not too long ago, small towns were the norm. The lady across the street from you tracked your every coming and going. She peeked out through her window and watched when you went to the doctor, and saw that you brought home a new TV, and knew who stayed with you over the weekend. But you also watched her through your window. You knew what she did on Thursday nights, and down at the corner drugstore you saw what she put in her basket. And there were mutual benefits from this mutual surveillance. If someone she did not recognize walked into your house when you were gone, she called the cops. And when she was gone, you picked up her mail from her mailbox. This small-town coveillance worked because it was symmetrical. You knew who was watching you. You knew what they did with the information. You could hold them accountable for its accuracy and use. And you got benefits for being watched. Finally, you watched your watchers under the same circumstances.
We tend to be uncomfortable being tracked today because we don’t know much about who is watching us. We don’t know what they know. We have no say in how the information is used. They are not accountable to correct it. They are filming us but we can’t film them. And the benefits for being watched are mur
ky and concealed. The relationship is unbalanced and asymmetrical.
Ubiquitous surveillance is inevitable. Since we cannot stop the system from tracking, we can only make the relationships more symmetrical. It’s a way of civilizing coveillance. This will take both technological fixes and new social norms. Science fiction author David Brin calls this the “Transparent Society,” which is also the name of his 1999 book summing up the idea. For a hint of how this scenario may be possible, consider Bitcoin, the decentralized open source currency described in Chapter 6 (Sharing). Bitcoin transparently logs every transaction in its economy in a public ledger, thereby making all financial transactions public. The validity of a transaction is verified by a coveillance of other users rather than the surveillance of central bank. For another example, traditional encryption used secret proprietary codes guarded closely. But a clever improvement called public key encryption (such as PGP) relies on code that anyone can inspect, including a public key, and therefore anyone can trust and verify. Neither of these innovations remedy existing asymmetries of knowledge; rather they demonstrate how it is possible to engineer systems that are powered by mutual vigilance.
In a coveillant society a sense of entitlement can emerge: Every person has a human right to access, and a right to benefit from, the data about themselves. But every right requires a duty, so every person has a human duty to respect the integrity of information, to share it responsibly, and to be watched by the watched.
The alternatives to coveillance are not promising. Outlawing the expansion of easy tracking will probably be as ineffectual as outlawing easy copying. I am a supporter of the whistle-blower Edward Snowden, who leaked tens of thousands of classified NSA files, revealing their role in secretly tracking citizens, primarily because I think the big sin of many governments, including the U.S., is lying about their tracking. Big governments are tracking us, but with no chance for symmetry. I applaud Snowden’s whistle-blowing not because I believe it will reduce tracking, but because it can increase transparency. If symmetry can be restored so we can track who is tracking, if we can hold the trackers accountable by law (there should be regulation) and responsible for accuracy, and if we can make the benefits obvious and relevant, then I suspect the expansion of tracking will be accepted.