U.S. Transportation Secretary Realizes Autonomous Cars Aren’t Ready
Despite the automotive industry having had the concept on its mind since the 1950s, autonomous vehicles still have yet to manifest in a manner that would allow them to be safely fielded in large numbers. With manufacturers previously vowing to have self-driving cars available to customers by 2020, consumers are starting to write the technology off as an industrial chimera. It’s also starting to look like the government is having doubts, especially now that U.S. Transportation Secretary Pete Buttigieg seems to be calling the technology to task.
“It feels like the widespread use of autonomous driving is seven years away, and it’s been seven years away for 10 years,” Buttigieg was quoted as saying by Jalopnik. “So the question is, will it be seven years away 10 years from now, or will we actually be getting somewhere?”
While your author has frequently complained about the Department of Transportation’s position in regard to widespread electrification, it’s nice to see someone within the government finally taking a realistic stance on self-driving vehicles. Though, with the automotive sector having promised the tech for years and seeing limited success, it’d be surprising to see anyone retaining a gung-ho attitude in 2022.
To be fair, the technology has made some meaningful headway since 2015. Numerous automakers have released advanced driving suites that can mimic certain aspects of autonomous driving. But these systems are often befuddled by inclement weather conditions.
Computers can be similarly confused by an unfamiliar environment or obstacle – things a human driver could instantly adapt to. This is why you can see companies like Waymo and Cruise building fleets of driverless cars that seem to have achieved success only to become the subject of criticism when they eventually go haywire. There are also legal hurdles and a mark of excellence that requires autonomous tech to perform perfectly to be of any real value to the industry.
It’s an incredibly high bar and one that’s starting to look totally out of reach – which is to say nothing about unanswered legal questions and the monumental cost automakers have incurred thus far just to develop this stuff.
However, this hasn’t prevented automakers from taking extra steps to implement safety measures in anticipation of self-driving vehicles. Hell, the entire premise of “mobility” revolved around building connected cars so that AVs could better interact with each other. By 2020, we were even seeing mainstream automakers adding driver-monitoring cameras to select vehicles that used advanced versions of cruise control – operating under the claim that such technologies will be helpful in preventing distracted driving while also helping put the legal onus back onto drivers. Will these invasive inclusions be scrapped now that it looks like self-driving cars are nowhere near ready to be used by the public?
Of course not. Automakers believe there’s a fortune to be made via in-car subscriptions and customer data harvesting. One could even make the argument that the industry claiming various systems would serve as an essential safety net for the eventuality of self-driving cars served as a Trojan Horse for a lot of the services automakers sees as more profitable. Whether or not that was the intent is debatable. But that’s still more-or-less the outcome.
Congress went along with the premise for years, often allowing companies to field autonomous test mules based on little more than an empty promise that they were following some kind of safety protocol. The Department of Transportation has been largely complicit in this and even embraced the promise of self-driving vehicles as a way to eliminate fatal car accidents altogether. Meanwhile, practically nobody seemed to notice that every modern vehicle was being outfitted with large, distracting screens and advanced driving aids that appear to be making everyone a worse driver.
Granted, this has changed in recent years. We’ve started to see the National Transportation Safety Board (NTSB) focus on car companies other than Tesla and groups like the American Automobile Association (AAA) have been testing just how fallible modern driving aids happen to be in the general sense. It’s taken nearly 10 years but word has gotten around. Sadly, some of the resulting actions feel like too little too late.
“There is a very serious danger right now in this kind of valley of death between where we started and where we’re headed, where these technologies do run the risk of making things worse,” Buttigieg continued. “Especially if people see ADAS, which is an automated driver assistance system, and treat it like a driver replacement system.”
“Just to be clear, I don’t care what they call these things, Autopilot or Self-Driving or whatever, there is no car that you can buy today from a dealer where you don’t have to be paying attention at all times when you’re driving.”
Though this all seems to exist counter to statements he made at roughly the same time. In an interview with Quartz discussing the alleged need to change people's psychology on all-electric vehicles, Buttigieg claimed something similar would need to be done to ensure autonomous widespread vehicle adoption could take place.
“Human drivers aren’t just problematic, they are murderous. Forty thousand people a year die in a car crash. And we have been bathed in this level of carnage all our lives. And so we’re a bit like people who grow up in a place that’s experiencing a war, in terms of how normal we think that is," he said. "[Technology] is not always the answer to everything. But frankly, it would be hard to do worse than human drivers when it comes to what we could get to theoretically with the right kind of safe autonomous driving."
“I think there is a sort of promised land on the other side. But psychology is going to be tough,” added Buttigieg. "[If] robots killed 10,000 people a year on the roads, there would be an uproar. But that would represent a 75 [percent] reduction in roadway deaths compared to where we are now. So we’ve got to make sure the reality and the perception of it is moving in the right direction.”
[Image: lev radin/Shutterstock]
Become a TTAC insider. Get the latest news, features, TTAC takes, and everything else that gets to the truth about cars first by subscribing to our newsletter.
Zerofoo on Nov 04, 2022
We don't know how the human brain works (mostly), yet we are trying to mimic that complex system in computer hardware that is millions of times less capable.
Anyone that knows anything about neural nets knows this whole exercise was doomed from the start.
I was on a team that built a very simple neural net in college. We successfully trained it to recognize the color blue. That took us an entire semester to get done and it mostly worked.
30 years later we are convinced we can train large scale matrices to perform complex decisions. Here's the truth - these systems are decent at pattern recognition, but they absolutely stink at learning and making decisions based on that learning.
Until we figure out how the human brain does this, we have no hope of emulating this in traditional computational systems.
Latest Car ReviewsRead more
Latest Product ReviewsRead more
- IH_Fever EV charger on a GM lot, probably with a Cummins generator to keep them running. A regular melting pot haha
- Tassos Wake me up when VW (or any other loser "Legacy" automaker comes up with a "BETTER TESLA" BEV AT THE SAME PRICE. SO far, VW has FAILED MISERABLY AND LOST BILLIONS DOING IT. Its models are way underwhelming and inferior, and cost not much less than the model 3. ANd DESPITE the SCANDALOUS $7,500 tax credit, which is an INVERSE ROBIN HOOD, takes from the average household and gives it to the average BEV buying family, which has an income of $170k+, VW STILL FAILED.ALso notice the so-called "Mobility Officers" at FORD AND Renault QUIT. another HUGE SCAM, Autonomous Vehicles, they wasted 100s of billions (all idiot legacy makers together) and predicted billions of profits, but so far they DROWN IN A SEA OF RED INK with NOTHING to show for it. Morons will be morons, and the ones in this forum will cheer for their failures "AWESOME, WV, Indeed"! LOL!!!
- Jwee More range and faster charging cannot be good news for the heavily indebted and distracted Musk.Tesla China is discounting their cars. Apart from the Model 3, no one is much buying Tesla's here in Europe. Other groups have already passed Tesla in Europe, where it was once dominant.Among manufacturers, 2021 EV sales:VW Group 25%, Stellantis at 14.5%,Tesla at 13.9%Hyundai-Kia at 11.2% Renault Group at 10.3%. Just 2 years ago, Tesla had a commanding 31.1% share of the European EV marketOuch. https://carsalesbase.com/european-sales-2021-ev/@lou_BC, carsalebase.com changed their data, so this is slightly different than last time I posted this, but same idea.
- Varezhka Given how long the Mitsubishi USA has been in red, that's a hard one. I mean, this company has been losing money in all regions *except* SE Asia and Oceania ever since they lost the commercial division to Daimler.I think the only reason we still have the brand is A) Mitsubishi conglomerate's pride won't allow it B) US still a source of large volume for the company, even if they lose money on each one and C) it cost too much money to pull out and no one wants to take responsibility. If I was the head of Mitsubishi's North American operation and retreat was not an option, I think my best bet would be to reduce overhead by replacing all the cars with rebadged Nissans built in Tennessee and Mexico.As much as I'd like to see the return of Triton, Pajero Sport (Montero Sport to you and me), and Delica I'm sure that's more nostalgia and grass is greener thing than anything else.
- Varezhka If there's one (small) downside to the dealer not being allowed to sell above MSRP, it's that now we get a lot of people signing up for the car with zero intention of keeping the car they bought. We end up with a lot of "lightly used" examples on sale for a huge mark-up, including those self-purchased by the dealerships themselves. I'm sure this is what we'll end up seeing with GR Corolla in Japan as well.This is also why the Land Cruiser has a 4 year waitlist in Japan (36K USD starting MSRP -> buy and immediately flip for 10, 20K more -> profit) I'm not sure if there's a good solution for this apart from setting the MSRP higher to match what the market allows, though this lottery system is probably as close as we can get.