Drivers, who needs 'em? It looks like some US government scientists are completely onboard with Uber and Google's plan to do away with drivers altogether. Fleets of autonomous taxis should vastly reduce the amount of overall emissions released in cities, and the economics of a driverless taxi already make sense today, according to a new Department of Energy study.
The study, published by Jeffery Greenblatt and Samveg Saxena of the Lawrence Berkeley National Laboratory in Nature Climate Change, posits that taxis are likely to be among the first industries to go fully autonomous.
Today's taxis are driven between three and six times as far annually as a standard passenger car, meaning that most of the cost of operating a taxi is in paying the driver, buying gas, and doing maintenance on the car—the initial cost of the car itself is much less important.
Here's the math used by Greenblatt and Saxena to argue that even today's expensive autonomous vehicle technology makes sense today.
"In New York City in 2005, only 24 percent of taxi fares went toward vehicle costs, with 57 percent going to drivers … driver income constitutes $97,600 per year, which could more than cover the incremental cost of autonomous vehicle technology [estimated at $150,000]. Even using current costs, if financed using identical model assumptions for vehicle capital, this would amount to $36,500 per year, 37 percent of New York City taxi driver income and 21 percent of total taxi fares. Therefore, autonomous taxis could replace current taxis at current autonomous vehicle costs and possibly even lower fares, providing an important early market niche." [emphasis mine]
Greenblatt and Saxena suggest that, given those potential cost savings, if the technology matures to a point where it's reliable, today's taxi drivers don't stand a chance. That is, of course, what taxi companies have been very much worried about as Uber makes inroads throughout the world.
Uber's CEO Travis Kalanick has repeatedly said he wants robots to replace drivers, and the company hired 40 engineers from Carnegie Mellon University earlier this year to develop driverless technology.
"As autonomous vehicle costs fall, it may become difficult for conventionally driven taxis to compete, and autonomous taxis may become ubiquitous," Greenblatt and Saxena wrote.
In an accompanying analysis article, Austin Brown of the National Renewable Energy Laboratory writes that it's not much of a stretch to imagine a near future where taxis are fully autonomous. He says that we are in fact being primed for such a future already.
"Consumers have already begun using analogous on-demand services such as Uber and Lyft, which could eventually begin offering autonomous options," he wrote.
Staggeringly, electric driverless taxis could push greenhouse gas emissions down by 94 percent on a per-vehicle basis (which was the initial purpose of the study), with emissions being completely eliminated if sustainable energy such as solar power is used to recharge them.
Autonomous taxis would be matched with riders and their destinations, meaning you'd essentially carpool everywhere, and the cars would be full more often than conventional taxis. The cars would autonomously recharge in between trips, keeping them on the road and reducing the total number of cars you'd need.
So this is the promise of autonomous taxis, if the technology continues to mature and if lawmakers (and passengers) are willing to let them shuttle people around city streets. There remain plenty of unanswered questions, however.
The estimated 233,000 taxi drivers out there will need to find new jobs, and taxi companies notoriously have a pretty strong local lobbying presence. Already, cabbie versus Uber tensions have flared up in cities all around the world. What happens when driving even for a ridesharing company is no longer an option? And what happens when Uber or Google control all of the taxi services in a given city?
If driverless technology continues to advance as it has (and come down in price), these are questions we may need to answer sooner rather than later.