Over 2.5k subscribers
For decades, we have been building our cars to save us from ourselves. As a result, we have made them heavier and less aerodynamic, driving up their cost. The smart (and more efficient) thing to do would be to take the driver out of the equation. Today, at long last, that might be feasible.
This is a link-enhanced version of an article that first appeared in the Mint. You can read the original here. For the full archive of all the Ex Machina articles, please visit the website.
Each year, over a million people die in automobile accidents around the world. While this might seem like a high number in absolute terms, it has been trending downwards on a per capita basis from its peak in the mid-20th century.
For the most part, this is because of the tremendous effort that manufacturers have put into making sure that their vehicles are safe. Since the vast majority of automobile accidents are on account of human error, today our cars come with airbags, crash harnesses, crumple zones and reinforced structures to give passengers the best chance of escaping unscathed. So much so that for more than half a century, our design efforts have been focused on protecting us from ourselves, rather than building more efficient transportation.
All this has added to the weight of the car, forcing manufacturers to make their engines even more powerful, which has, in turn, further increased the
There is an economic cost to all this. Since they are designed to meet crash standards, our cars today are far less aerodynamic than they could have been, and, as a result, consume far more fuel than they need to. Given the additional weight of their safety features and much larger engines, their tyres suffer far more wear and tear, which has increased maintenance costs all around.
We know that it is technically possible to design cars up to ten times lighter without sacrificing speed or comfort. All we need to do is adopt a more aerodynamic design and a more lightweight frame—one that simply protects occupants from the elements and nothing more. This, however, is currently unthinkable because there is no way to ensure that light aerodynamic cars are safe.
Or is there? Since 95% of all accidents are caused by human error, surely the most effective way to reduce automobile fatalities would be to take humans out of the equation. A decade ago, this might have been idle speculation, but over the last ten years, autonomous vehicle rental companies like Waymo have been operating truly driverless vehicles on public streets in the US with much success. As a result, we now have evidence to show that such vehicles are far safer than those with human drivers behind the wheel. With reaction times 150 times faster than a human’s, they cause 88% less property damage and 92% less bodily injury. There is no reason not to use them more widely.
Without a human at the wheel, cars will no longer need to be built like rolling fortresses. They can be smaller, lighter and more aerodynamic without the overpowered engines we currently use to transport the safety features we can’t do without. If we deploy them as shared fleets, a single vehicle will replace dozens of private cars, increasing utilisation and dramatically cutting fuel use per passenger-kilometre covered.
While this might seem like some futuristic fantasy, in some parts of the world, it is already a reality. In August 2023, Waymo was running 10,000 paid rides a week. By mid-2025, that has grown to over 250,000. In under two years, cumulative trips grew from 1 million to over 10 million. The average Waymo cab now completes more daily trips than 99% of all human Uber drivers. We’ve seen this S-curve before—in smartphones, mobile data and digital payments with India’s UPI—and we know that once it starts, it accelerates rapidly.
If this is the future, India should not be left behind. But can we even hope to go down this path? It’s all well and good for us to talk about fleets of autonomous vehicles in the well-ordered streets of San Francisco. It’s another thing altogether to imagine something like this on India’s chaotic streets.
While this might, at first, seem like an insurmountable problem, traffic is a system, and systems adapt. Autonomous vehicles follow the rules strictly. They don’t jump signals, block intersections or squeeze through impossible gaps. If forced to deal with humans who do all that and more, they may be rendered immobile, unable to proceed in any direction, given the chaos that their sensors perceive. But as their share of the traffic mix grows, the predictable behaviour of driverless vehicles will start shaping the behaviour of human drivers around them. Once self-driven vehicles are in the majority, it is the erratic human drivers who will be compelled to comply.
While the transition to autonomous might be difficult for countries where a large proportion of the population already owns cars, this should not be a concern in India, where that number is less than 10%. Since so many of us rely on public transport, India’s transition to autonomous vehicles can start there. Driverless metro trains are being tested. Many cities have dedicated bus lanes, and it should be possible to insinuate autonomous vehicles into those corridors with little disruption to normal traffic. As drivers and pedestrians get accustomed to them, we could add cars and other similar vehicles.
For decades, we have purposely built our vehicles to save us from ourselves. Autonomous vehicles give us a chance to break that pattern, allowing us to redesign this mode of transport from the ground up: lighter vehicles, higher utilisation, less fuel, more order.
In a country that imports most of its oil, that’s not just good engineering. It’s critical efficiency.
Rahul Matthan
Support dialog
