Aauthored by Saeculum Research
- Beware: Driverless cars will not happen for another two decades, at a minimum. In recent months, automakers and tech firms have been racing to bring driverless cars into commercial production. While they claim fully autonomous vehicles will be on the road by 2018 (Tesla) or by 2021 (BMW), this timetable is wildly optimistic. The jump from semiautonomous to fully autonomous is vastly more challenging than the industry realizes. Why? Semiautonomous features have already mastered the easy tasks. Full autonomy requires staying on the road, making sense of complex situations, interpreting disorienting changes in light, and much more—tasks that remain well beyond the power of today’s sensors and AI. Maybe this is a good thing: We may need more generational turnover before attitudes and behaviors can fully adjust to the driverless paradigm.
- Watch out for backlash against semiautonomous automakers. The race to fully autonomous vehicles is an all-or-nothing bet: Either people can take their attention away from the road or they can’t. While Elon Musk claims that fully autonomous vehicles will be available in 2018, Tesla’s reputation will take a beating if he can’t back it up. This realization will take the luster off of luxury brands like Audi and Cadillac that are currently testing out semiautonomous features before they are adopted by lower-tiered brands. If lawmakers rule that cars must either be self-driven or fully autonomous for safety purposes, luxury brands may be the hardest hit.
- Don’t believe any hype about “cars without steering wheels.” While Google is currently complying with California regulations that mandate human controls, the company is determined to create a car that does not require a steering wheel or pedals. Although this might work for specialized commercial purposes (like in mining and forestry), this concept will never migrate from the testing ground to the sales floor. To accommodate such vehicles, the United States would need an infrastructure overhaul. At the very least, there would also have to be some sort of override in the event of an emergency.
- Don’t expect technology alone to create the 100 percent driverless car, even in the long run. While better integration of satellite and LiDAR and the growth of vehicle-to-vehicle technologies will certainly help semiautonomous vehicles nudge closer to autonomy, the perfection of driverless cars requires more than just extra tech in the car. We will certainly need a new and improved highway system designed for a driverless society. We will probably also need a wireless communications infrastructure that allows all cars (even those with drivers) to signal to each other. Such efforts will require massive public investment. We will not get there only with new and improved gadgets added to each car.
On May 7th, 2016, a man died after his Tesla crashed into a tractor-trailer while in autopilot mode. According to witnesses, the driver was watching a movie and never applied his brakes. This is unwelcome news at a time when the race to produce fully autonomous vehicles is being breathlessly heralded by the media. But it is a useful if tragic corrective to the overhyped optimism about how Silicon Valley and Detroit will make driverless cars commonplace within a few years.
In fact, their widespread use is still decades away. Standing in the way are not just vast cost, legal, and security obstacles, but also the fundamental limitations of AI technology. The journey from semiautonomous to fully autonomous is actually much more difficult than most people realize. And it may require decades of generational turnover before most consumers are comfortable with a driverless transportation paradigm.
Autonomous vehicles, along with various autopilot add-ons for ordinary vehicles, have been under development for well over a decade. But in recent years—and especially over the last few months—the race to bring them into commercial production has accelerated. In March, GM bought self-driving startup Cruise Automation for $1 billion. In May, Toyota and Uber joined forces, Apple invested $1 billion into Chinese ride-sharing company Didi Chuxing, and Google partnered with Fiat Chrysler—all in the pursuit of self-driving vehicles. These players are betting heavily on transforming the $5.4 trillion transportation services market (buses, cabs, and passenger rail). While Tesla proclaims that it will have autonomous vehicles commercially available by 2018, BMW is planning for 2021—leaving many to believe this goal will be achieved in five years.
This timetable is unrealistic. In fact, the widespread use of driverless cars will not happen for at least another 20 years. Some of the likely causes of delay are already well understood. These cars are still too expensive. (Today’s Google car costs six figures to build.) They will require a whole new legal and insurance infrastructure. (After an accident with these cars, you sue the car company, not the other driver.) And they are unnervingly vulnerable to hacking (or mass hacking, which could make this a national security issue).
Yet driverless cars face a more intractable obstacle: the extreme difficulty of moving from semiautonomous to fully autonomous. Hype has gotten ahead of reality thanks to the rapid adoption of semiautonomous features like lane departure warnings and automatic parking. (Tesla and Audi use terms like “piloted” and “assisted” driving.) Consumers naturally assume that the transition to full autonomy is an easy next step. According to many experts, however, this is a grave error. David Mindell, renowned authority on engineering and automation, believes totally autonomous vehicles are a false hope. “I’m not arguing this from first principles. There are 40 years’ worth of examples.”
Why is it so much harder to go that last stretch? Most of today’s semiautonomous features take over the most routine driving tasks. All higher-order thinking remains with the driver. And what will it take to replicate that higher-order thinking? Ah, there’s the rub.
Here are some critical difficulties that will challenge any version of driverless car likely to be available in the near future.
First, staying on the road or lane in all circumstances. Driverless AI cannot cope with road edges or lane markings obscured by water, snow, gravel, or dirt—or just lane markings weathered by age or incorrectly veering off the road. And let’s not forget construction zones and detours, where human intelligence is required to decide where the car should go. Yes, driverless cars have GPS, but none will dare to abandon its LiDAR or camera for accurate perception of the immediate environment. Bad weather is a big problem here: In rain or snow, current autonomous cars are programmed to stop. Yes, that’s right, they just stop.
Second, making sense out of confusing and ambiguous situations. Think of parking lots, frontage roads, and toll plazas. Here, the lines disappear and judgement takes over. Think of obscured or ambiguous signage. Or a police officer signaling traffic. Humans can figure out what to do when the rules are unclear, provisional, or absent. It’s harder for a driverless car to understand that it’s sometimes acceptable to drive in the opposite lane.
Third, interpreting disorienting changes in light. Today’s sensors can’t recognize the color of the traffic light through the sun’s glare, know if the approaching dark spot on the road is a shadow or a pothole, or determine whether a puddle is shallow enough to drive through or deep enough to take out a tire. These judgments have life-threatening consequences. In the recent Tesla accident, the autopilot sensors couldn’t distinguish between the white side of the tractor-trailer and the brightly lit sky.
Fourth, engaging in higher-order object recognition. Humans know that it’s safe to drive through crumpled paper or plastic bags. They know to swerve around tire debris and pieces of fender. How is a computer supposed know the difference and act accordingly? Humans are also able to figure out which signal in a complicated string of traffic lights belongs to their lane. While tech enthusiasts assume that GPS would give driverless cars updated information on traffic lights and stop signs, GPS simply isn’t accurate enough to distinguish between adjacent lights or signs.
Fifth, comprehending human intentionality. Only a human can easily assess the intentions of other humans, whether it’s dealing with road-rage drivers or dodging cars taking evasive maneuvers from emergency vehicles. It takes a human to “read” the directions of a police officer, the gestures of a driver in distress, or the hesitation of a deer in the road. When a ball rolls into the road, only a human can sense when a child is likely to be following.
The lesson to be learned: Replacing the judgment of the driver becomes exponentially more difficult the further you go. Replacing 100 percent of the judgment of the driver is almost certainly impossible. And that’s the problem: At anything short of 100 percent, drivers will need to be completely attentive—because they may have to take over at any moment. This already poses an awkward tension in the marketing for current semiautonomous features (like those on a Tesla or Audi). The company tells you that this feature will allow you not to pay attention. But then they tell you always to pay attention.
In the long run, to be sure, programmers and carmakers will certainly achieve their goal of producing an affordable driverless vehicle fit for widespread use. But it’s not going to happen on anything close to the timetable that these companies and the industry media are now touting. It will require decades.
This slow pace may be necessary in any case to overcome one more obstacle on the road to driverless transportation: generational inertia. While Boomers and Xers have spent a lifetime wedded to the concept of a car as a projection of personal autonomy, Millennials are more comfortable ceding personal control to an intelligent system and are less attracted to ownership and control. This vision, however, may only be possible in a world in which most of society is younger than Millennials, not older—meaning that substantial generational turnover may need to happen first.
Material posted on this website is for informational purposes only and does not constitute a legal opinion or medical advice. Contact your legal representative or medical professional for information specific to your legal or medical needs.