Designing for Natural Interfaces: Mobility – Lecture notes

April 18th
8 mins read

Designing for Natural Interfaces: Mobility – Lecture notes

In 2007, California’s DARPA Urban Challenge tasked eleven ‘robot’ cars with safely driving sixty miles alongside human-controlled vehicles—entirely free from people keeping things under control behind the wheel. It was the first competition of its kind, and thanks to robust human programming, all of the cars were able to complete the task to ample praise. But in retrospect, calling them ‘robot cars’ back then was probably a stretch. Had this competition happened a mere eleven years later, there’d be no need for people to spend over a year meticulously programming driverless cars in order to pull it off. Thanks to advancements in processing power and machine learning, semi-autonomous vehicles are already driving worldwide. In fact, they’re often anticipated to saturate markets by 2022—and multiple countries are striving towards bringing fully autonomous vehicles to city streets already within the next two or three years.

Speeding Towards the Future

Take China, where autonomous cars are key to ‘Made in China 2025’, a nationwide program that’s meant to transform the country into a world leader in innovation. That’s helped tech conglomerate Baidu develop Apollo, a self-driving software that’s been tested on city roads since last year; meanwhile, startup UISEE is ready to roll out an autonomous vehicle, designed to operate in a closed environment with fixed roads and predictable circumstances. China’s innovation in this field will only continue to accelerate thanks to the Chinese Ministry of Industry and Information Technology’s new guidelines: released just last month, they make it easy for local authorities to road test autonomous vehicles. Already, the ministry has issued licenses for self-driving cars from two companies (electric car start-up NIO and state-owned giant SAIC Motor.)

Meanwhile in the US, some experts predict that fully autonomous vehicles will be broadly deployed by 2020; and last September, the American House of Representatives unanimously approved a bill to quickly allow self-driving cars without a human controller on roads. Because monoliths such as Mercedes, General Motors and even Toyota are all scrambling to bring self-driving cars to market – presumably to avoid becoming irrelevant when oil-run cars could become vastly outnumbered by autonomous, electric cars by 2030 – the lobbying from these companies has presumably been fierce enough to triumph over safety concerns from citizens about the maturity of this technology. (Even Uber, who recently suffered a PR disaster when a semi-autonomous vehicle was involved in a fatal collision, remain firm in their resolve to invest in self-driving cars.) In Germany, on the other hand, Audi are trying to commercialize their new A8 which would make them the first automobile company to release a Level 3 autonomous car (enabling the driver to completely take their hands off the wheel and the car to accelerate and slow down without assistance.) And starting in 2019, France will have the legislative framework to allow Level 5 autonomous vehicles on the road—meaning, 100 percent autonomous, self-driving cars, no human control necessary.

On the Racecourse

Indeed, it seems like everyone wants a piece of the self-driving car pie—but it’s no wonder when you look at what’s at stake. For right now, we don’t have enough use cases to precisely anticipate how autonomous cars will affect society; the technology hasn’t come far enough for these vehicles to be rolled out en masse, much less the legislation to make that a viable reality. So today, fully autonomous vehicles remain a mysterious, fruitful question mark—which is why so many companies and governments are in a race to bring the best autonomous driving technology to market. Since no one’s drawn the golden ticket of autonomous technology, those who do so first will presumably influence people’s lives in intimate and unparalleled ways.

The Headlight at the End of the Tunnel

Admittedly, that ‘car race’ may sound like fertile ground for a dystopia of ever-present nanny states that use your car to control you. But it’s important to remember that almost no technological development is that black and white—so when it comes to self-driving cars, they could be responsible for vast benefits to our lives. For example, an overarching dialogue surrounding autonomous vehicles has to do with road safety. Right now, over 90 percent of road accidents are caused by human error; in the US, a third of fatal accidents in 2015 involved alcohol. Autonomous vehicles could drastically reduce the rate of road collisions by eliminating the human error factor (and freeing up the driver to enjoy an alcoholic beverage in the car, as the wine industry apparently hopes.) Cycling would become safer, which could impact public health; cars would idle less and therefore produce 20 to 50 percent less carbon dioxide, which could do wonders for the environment. Even congestion could be alleviated: a car that’s free of human error would be more efficient on the roads, which could lead to less traffic, lower commute times and even a better parking model.

Aside from making roads safer and more efficient, autonomous cars could increase mobility for low-income communities or people who have trouble driving (think the elderly or those with chronic injuries.) On average, these groups spend 16 percent of their income on transportation and have to navigate an inconsistent public transport model (anecdotal evidence shows that buses often will reroute away from ‘dangerous’ neighborhoods or ditch service to areas only a few people need to get to.) And since low-income people often live on the outskirts of cities, their commute times are longer, which is associated with less chances of moving up the economic ladder. In contrast, a 2015 study from Harvard found that access to efficient and affordable transportation is associated with upward mobility in employment, education and healthcare. By 2030, autonomous technology is meant to lower costs-per-mile as well as the costs of maintaining efficient public transportation; some even speculate that robo-taxis could be subsidized in the future. Following that logic, self-driving cars could offer new options for low-income populations to get to where they need to go, which would ultimately increase their social opportunities.

What’s a Car, Anyway?

But perhaps the most intriguing idea hovering around self-driving cars is the complete transformation of the very concept of a car. For if we don’t have to spend our time in cars actually driving, what could we be doing instead? Perhaps we could work in a sort of mobile office, enabling us to spend less time in our ‘actual’ office. Perhaps we could use these cars as hotel rooms when we travel and have them shuttle us to our destinations while we sleep or eat a meal. (No more having to navigate the jungle of AirBnB or hotel websites to find a place to rest your head at night.) If we stop thinking of cars as vehicles and instead view them as “driverless rooms sitting atop an all-electric drivetrain”, the opportunities start to seem endless—both for people experiencing these services, but also for brands providing them for us.

Granted, that somewhat utopian vision of a life on wheels is quite a few years away. From a governmental standpoint, the more tangible benefits of self-driving cars have to do with saving lives and being environmentally conscious. More pressingly, though, there’s economic incentive. For example, the US government estimates that $240 billion of government funds, taxes and insurance dollars are spent a year on road accidents; considering that US car sales in 2016 amounted to roughly $600 billion, that’s over a third of lost profits. (A separate analysis found similar rates of profit loss due to accidents in the UK.) Add the promise of an easier way to hand out traffic tickets (if autonomous cars will be programmed to report on their users’ traffic violations, as some writers speculate) and you’ve got an appealing formula for governments to reap the benefits of automation.

I Spy With My Little… Vehicle?

Which brings us to the more troubling implications of a society full of autonomous vehicles. After all, these cars safely drive because of a combination of computer vision-enabled cameras, sensors and machine learning algorithms. Machine learning works by exposing these cars to mass amounts of data about roads and behavior so that they can learn to drive safely; computer vision lets them monitor and understand their surroundings so that they can react accordingly to situations. That combination begs the question: what kind of data do these cars absorb about the passenger and their surroundings—and who do they feed it to? John Krafick, CEO of Waymo aka Google’s Self-Driving project, recently said data collection ‘is not a priority’ but eventually admitted that it could become a ‘workstream’ in the future. “I don’t believe a word he is saying,” proclaimed one journalist. “Google is in the market of data harvesting, and all the better if a driver is ‘trapped’ in a car, a captive market for advertisers.”

Besides autonomous vehicles potentially functioning as “panopticons on wheels”, some feel the self-driving utopia could actually turn into its warped mirror image. An office on wheels won’t give us more free time, they say; it will only increase demands to work even while commuting. Marginalized communities won’t benefit; instead, that’s where self-driving cars will be stored when not in use to stop them from becoming an eyesore for more privileged people. Autonomous vehicles won’t free up congestion and make for more livable cities; instead, we’ll be at the mercy of technologists who “imagine the driverless world with a distinct lack of imagination”. We’ll no longer have the luxury of looking out the window at the world outside; instead, those windows will become moving ads—a sort of eBay on steroids we can’t escape from.

Get Behind the Wheel

In short, the discussion around self-driving cars seems supercharged by two competing forces—blind hopefulness and fear-mongering sensationalism. Which is too bad: ultimately, this lack of nuance disempowers citizens from participating in the discussion and encourages a passive “wait and see what happens” approach. For although autonomous vehicles will increasingly be adopted, it’s not going to happen overnight; as Bryan Salesky, CEO of Argo AI, says, “those who think fully self-driving vehicles will be ubiquitous on city streets months from now or even in a few years are not well connected to the state of the art.” That means people, not just companies, have time to take a role in defining the future of self-driving cars. Programs like the Driverless City Project – an interdisciplinary research project that takes a playful, rigorous approach to envisioning the fully autonomous future – are already beginning to encourage that kind of engagement, but we need more of these types of initiatives. Perhaps that looks like cities demanding that autonomous vehicle manufacturers provide transparency around data and safety before they’re allowed to carry out road tests; perhaps it’s “fiction writers, monks, rabbis, architects and urban planners” banding together to prototype the self-driving future.

Whatever the solution is, there’s little doubt that we’re speeding towards a life on wheels. If we want to avoid surrendering it to big tech or government, let’s make sure we’re sitting in the driver’s seat from the beginning.