The Future Is Machine-Readable

Robots and automated cars will require our cities to become machine-readable, and soon.

A demonstration of the Dustbot in Sweden, 2009.

This article arises from Future Tense, a collaboration among Arizona State University, the New America Foundation, and Slate. Future Tense explores the ways emerging technologies affect society, policy, and culture. To read more, visit the Future Tense blog and the Future Tense homepage. You can also follow us on Twitter.

If I asked you to picture a city of the future, you might imagine some gleaming metropolis in Japan or Korea, a Masamune Shirow fantasy of plate glass and chrome, with driverless cars whooshing along highways.

But the real model for tomorrow’s city is a tiny medieval town in the Tuscan countryside called Peccioli. With a quarter of the town’s residents above pensionable age, even the populace seems antiquated. Yet Peccioli is at the forefront of Europe’s technological revolution. A collaboration with the Sant’Anna School of Advanced Studies of Pisa that began in 1995 made the town a testing ground for advanced research into aging, telepresence, alternative energy, environmental protection, and more. It was here that free-roaming robots first began working alongside the public.

After a demonstration in 2009, a pair of robots named DustCart took over municipal waste collection in the heart of Peccioli for two months in 2010. After residents called to request a pickup, the robot’s supervisor (an artificial intelligence named AmI) would dispatch the nearest DustCart to the client, whose rubbish would be collected and carried to a sorting station. The locals called the robots Oscar (a reference to a man who’d provided similar on-call garbage collection before, not the garbage-happy Muppet).

It might seem disappointing that in the 21st century, our best consumer robots are ambling waste collectors like Roombas and DustCarts. Why has the rise of the robots not been as meteoric as expected? While machines can easily outperform humans when it comes to simple, repetitive tasks, they have great difficulty operating in novel environments. These limitations kept them chained to the monotonous predictability of the production line and, for the most part, out of public sight. When free-roaming machines finally arrived, they took first to the air—the arena in which they encountered the fewest obstacles.

If robots are to become an everyday presence, the usual thinking goes, they’ll have to be able to function in a completely uncontrolled environment. However, I think the inverse is likely to be true: In the future, we will sculpt our environment to become more robot-centric to accommodate their needs.

Take, for instance, today’s highways, which would be utterly baffling to a robot. Google’s famed automated vehicle can only navigate itself after a human has programmed in “like road signs, traffic lights, and so on,” says Google’s Jay Nanncarrow. The Google car cannot operate in a novel environment, and it’s not clear how the system would handle unexpected events such as temporary traffic lights or lane closures.

If automated cars are really on the horizon, then we will have to invest heavily in infrastructure to make the roads robot-friendly. This could mean radio beacons at pedestrian crossings, or road markings inlaid with inductive loops so that cars can sense where to stop at complicated junctions. We’d also need signs to warn drivers as they leave a road that is configured to support automated cars.

On the street, shops and businesses might supplement eye-catching signs with ones suited to electronic optics. In Korea, commuters can now shop at virtual supermarkets by scanning murals of groceries plastered across metro platforms. The advent of inexpensive robotic systems might even herald a return to the old system of grocery shopping, where customers presented a list of goods for the store assistant to fetch.

Such a system has already been reborn at North Carolina State University’s new $115 million James B. Hunt Jr. Library. The school heralds it as “a symbol of the next wave of development” for the institution. One novelty of this new wave is that students will no longer be allowed to wander the stacks, as generations did before them. Two million volumes will be packed into a climate-controlled chamber underground, accessed only via a robotic crane. The new stacks occupy a fraction of the space previously required and recall will be much faster. (The system may also put an end to the age-old practice of students hiding in-demand books from classmates.) Rather than design a robot that could navigate the library, NCSU designed a library around the robot.

How will this balancing act manifest itself when robots migrate into other walks of life? Making highways friendly to self-driving cars will be relatively easy, as these roads are already a tightly controlled environment with a strong degree of standardization, from the layout of roads to their signage. Highways are often under the control of a central authority, too. This makes it relatively easy to implement any changes needed, so we shouldn’t be surprised if machine-readable road architecture becomes the norm. In Peccioli, the researchers added several features to the roads to support DustCart. A special “robot lane” was demarcated in yellow paint, to segregate the trundling robot from normal traffic and avoid congestion in the narrow streets. Signs were also posted to warn drivers of the unusual road user. The area was saturated with high speed Wi-Fi and CCTV to ensure the control station remained in constant contact with the drone, and navigational beacons were installed throughout the collection area to help DustCart find its way.

But the ascendance of robots is not just about building a robot-friendly controlled environment; it’s also about making as much of our surroundings as machine-readable as possible. QR codes, the information-dense mutant offspring of barcodes, are not a new invention, but they are finally coming to prominence thanks to the widespread adoption of smartphones. Rather than design systems with optical character recognition on par with our own, it proved easier to augment or sometimes even replace written text with machine-readable patterns that are incomprehensible to humans. It’s unlikely that we’ll give over our public information entirely to checkerboard symbols, but if we expect androids to pick up our shopping and transport patients across hospitals, at the very least we should expect this kind of bilingual signage to increase.

We are in the midst of a mass digitization of our data. We expect a huge proportion of our daily information, from news to correspondence, to be published in a format that can be delivered digitally. Meanwhile, archivists are busy scanning and uploading civilization’s back catalog of printed materials. But the future offers to take us one step further, where the physical nature of the built world itself will be machine-readable by default.