There’s this argument online that flares up in Facebook Groups and comment sections from time to time, positing that the advent and widespread use of driver assistance systems are making people worse drivers. Because many new cars will alert you if there’s a car in your blind spot, actively steer themselves to keep you in your lane, and even slam on the brakes if you’re about to rear-end someone, why should you have to be constantly worrying about those things?
Studies have shown that vehicles with tech like adaptive cruise control and automatic emergency braking might actually be doing more harm than good, as my colleague Jason covered last year. These systems do well when they work, but sew the seeds for more incidents when they don’t, since they allow driver focus to shift away from the active task of driving.
These systems are also making people faster drivers, according to a study recently published by the Massachusetts Institute of Technology. Additionally, the study determined that people who go from driving gas-powered cars to electric vehicles also tend to drive faster in some scenarios. Is all this new tech emboldening people to go faster? Sure seems like maybe.
The study, conducted in a collaboration between MIT’s AgeLab and Center for Transportation and Logistics and Toyota, analyzed driving data for 24 people in New England, who each drove a Tesla Model 3, an EV, and a Cadillac CT5, a gas-powered car. Both cars were equipped with their respective manufacturer’s highest level of driver assistance at the time: The Tesla had Autopilot, and the Cadillac had GM’s Super Cruise. Both of these systems allow for full hands-off driving, but still require the driver to pay full attention to the road.
Each of the 24 drivers had their driving habits analyzed for a month, resulting in 38,000 miles of total data, according to the study. Researchers used GPS tracking data, indicated speed, g-forces, and whether the driver assistance tech was active to determine driving habits and identify the differences from car to car.
The study found that when drivers used partial automation systems to assist their travels, they drove faster than those driving without the systems activated, not only for longer periods of time, but at higher speeds. From the study’s abstract:
[D]riving with partial automation, regardless of powertrain, was associated with significantly longer speeding durations and slightly greater speeding magnitudes compared to manual driving.
Interestingly, when you actually read the study’s conclusion, it doesn’t concretely tie automation use to faster driving:
[T]he findings indicate that regardless of powertrain, the probability of speeding appears to increase with automation use (Haus et al., 2022), which may or may not be a result of the presence of the automation software.
This makes a bit of sense, considering the study’s exceedingly small sample size for both the drivers and the number of miles driven. But I wouldn’t be surprised if the results held as the sample size increased. When you let the car do most of the work, and you become less of an active participant, you become less concerned—or even less aware—of stuff like speed, because you’re taking a back seat to the act of driving, rather than devoting your full attention to it. So you might not realize you’re going way faster than you should be.
When it came to EVs, the study found that, after some time behind the wheel, drivers tended to have more of a lead foot, but only on residential streets and controlled access roads (interstates and divided highways), and not arterial roads (high-volume, main roadways with multiple lanes that aren’t divided highways with controlled access points). From the study’s conclusion:
Compared to ICE drivers, the EV drivers tended to speed for less extended periods on arterial roads and were more likely to exceed speed limits by greater margins on residential and controlled access roads. Differences in speeding duration between EV and ICE drivers on arterial roads were evident from the start of the study. However, differences in speeding magnitude on residential and controlled access roads emerged after the first week, reflecting a gradual shift in speed control among EV drivers.
The delay in speed gains over time is the most fascinating point here, as it suggests that once people were comfortable with how the EV drove, they tended to gain more speed. Whether that means they were more confident about speeding or they couldn’t as readily perceive how quickly they were going is unclear.
Having switched between EVs and gas-powered cars in the past, I can say that in most cases, it’s far more straightforward to quickly bolt up to speed with an EV, and way easier to accelerate, even when you’re already at highway speeds, to pass people. There’s also no sound involved with an EV, meaning it’s tougher to immediately perceive how quickly you’re traveling. I could see the average driver making good use of that instant torque to get places a bit quicker, without realizing they’re doing it.
It’s worth noting the study was conducted not only to find out whether EVs or automation makes people faster, but to find out how people actually use their vehicles with these new pieces of tech to inform future car design, according to the Detroit Free Press:
Pnina Gershon, a research scientist at MIT’s AgeLab and Center for Transportation and Logistics who conducted the research, said the point of the study was to simply observe how people realistically drive the vehicles in their daily lives, not solve a specific engineering problem.
“There is limited data on this topic of how the introduction of new technologies are associated with changes in behavior,” Gershon said. “There is a gap here. A lot of attention is given to greenhouse gas emissions and so forth, but driver behavior is another key aspect that needs to get the spotlight.”
So how exactly will this study influence the next generation of Toyotas? Derek Caveney, the Toyota senior engineer and liaison between the company and MIT, didn’t really get into specifics when asked by the Free Press, saying that this study and those like it are “creating more opportunities to develop interventions that support safe and efficient driving behaviors.” To me, it sounds like the company wants to make sure people are more aware of their situations, even if they’re not in total control of their cars 100% of the time.
If anything, I think more research is needed on this subject. As automated systems become more commonplace, it won’t be long until the majority of cars on the road use at least some of this driver assistance tech. Whether that’s better or worse for everyone on the road, well, I guess we’ll see.
Top graphic image: DepositPhotos.com, Volkswagen, Jason Torchinsky
What “causes” me to speed is driving a fast car.
When I drove a TDI, It caused me to hyper-mile.
When I drove an Oldsmobile, it caused nothing as I knew I was driving an appliance.
In Soviet Russia, car drives you.
> Both of these systems allow for full hands-off driving
I don’t understand how those are even legal. You’re supposed to pay attention and ready to take over, yeah? We know how much braking distance increases for every extra second your reflexes take. Having to move your hands onto the wheel is going to take enough fractions of a second to make a difference in outcomes in the case of an adverse situation.
Until we have actual self-driving cars I don’t think those half-measures should be allowed on the market at all.
The solution is so easy! A GPS speed governor.
“oh but it might be inaccurate!”
I’d be fine with capping the limit at the maximum speed allowed for the state. If the fastest that state allows is 70 mph then that’s the most the governer permits.
“But you don’t know why that person is speeding! It might be an emergency!”
I’d also be fine with the car nagging the shit out of anyone driving anything over 5 mph of the posted speed limit for more than a few minutes and if the nagging isn’t heeded calling the cops while also recording and automatically submitting court admissible evidence of the speeding. If it’s truly an emergency the driver may get help, if not they can explain themselves to the court and their insurance.
N.O.
Y.E.S.
Do you want a police state? Because this is very much a path to a police state.
You prefer anarchy?
Keep in mind driving is a privilege, not a right. Enforcing the speed limit is not a violation of anyone’s rights.
I’m not too concerned about the speeding or lack thereof. I’m concerned about government-mandated devices that track and/or control vehicles. In a perfect world that wouldn’t be a problem, but our world is far from perfect.
If one is so concerned the best, easiest way to lower the justification for such devices is not speeding and actively discouraging others from speeding.
> But you don’t know why that person is speeding! It might be an emergency
Easily solved! The governor is on all the time when you’re letting the car drive. If you want to go faster, you have to control the steering, pedals, etc yourself.
Not sure how that solves anything other than becoming damning evidence of speeding if it ever went to court.
I was being facetious 🙂
Yeah, this surprises me too. I use my Comma just like cruise control. I set it a couple mph more than the speed limit and let it do it’s thing. It seems to give me more bandwidth to pay attention to the things happening around me. If I drive more aggressively, it’s in the Miat where I am shifting and enjoying the engine sounds.
Interesting. I use adaptive cruise in my Audi RS5 regularly to make sure I’m NOT going too fast. It’s really easy go faster than intended, but if I set cruise to the pretty ‘normal’ traffic speed, I can make sure the only time I go past that is when I do so intentionally to pass someone. Of course, it’s not a FSD by any stretch. It beeps at me if I don’t have steering input in about a minute…
The only times I’ve used cruise control (never own a car with it) has been for average speed zones, and I feel that it lets me give more attention to the road, cars around me etc, because I’m not staring at my speedo trying not to get a ticket.
Should be “sow the seeds” as in sowing and reaping.
Also, CT6, not CT5.
And now for the Tears for Fears hit Sewing the Seams of Love!
How many seeds could a seamstress sew if a seamstress could sew seeds?
Data shows a huge difference is rear end crashes for cars that have automatic emergency braking. That is not in doubt the tech works as intended.
This gets murky when we talk about people misusing Level 2 systems like FSD.
As to people driving faster in an EV – it is the lack of noise. I us my cruise control a lot more in my EV than my gas car because I look down and see I’m going faster than I intended.
(Same thing happened decades ago when I first drove my dad’s Oldsmobile 98 vs my old S15 Jimmy. The quiet and soft suspension was sneaky and I would look down to see I was going 80 mph (in the 65 mph days)
Yeah, it’s way easier to unintentionally speed in something with ample power and little to no NVH relative to what you are used to. EVs are the poster child for this, but in reality it happens almost any time you move upmarket or even just into a significantly newer vehicle.
I don’t have any autonomous systems in my car, but I do have adaptive cruise control and it’s a mixed bag.
I love automatic following distance. I open up to the maximum gap, set for some comfortable velocity usually a mile or two below the speed limit because I don’t need the added stress of trying to watch for state troopers, and relax.
I hate tailgaters. HATE them with the heat of a thousand suns. If I could equip my car with rear-facing shit cannons, I’d be getting a lot of use out of those. But the cruise control will not be cowed. It won’t be forced to go faster by someone climbing up my ass. I just keep it between the lines and relax.
The mixed part is the automatic lanekeeping assistance. I hate that almost as much as I hate tailgaters. It crowds the left side of the lane, which is completely unnecessary. If it held the middle that would be great, but it just won’t. It wants to ride the left line, and I’m not having it.
ADAS:
Advanced
Driver
Annoyance
Systems.
Active
DUI
Avoidance
System
Lets face it, the only people that these people help are people who shouldn’t be driving in the first place.
I mean, better than the likely alternative, but you’re not wrong. I do kinda wonder if lane departure / lane keep might’ve prevented me from going off the road after falling asleep. But then again, wrecking a car that I rather liked is a great way to make sure you learn a lesson and don’t forget it.
So all the nanny things I turn off so I can drive in comfort keep me from speeding?
That’s funny.
I dont need any of that getting in my way of actually driving. The only really useful parts are the parking sensors so I don’t bump a wall in the parking lot and the backup camera so I can watch the ant hill behind my car at home. I don’t want to run over the skitter critters. For everything else I have eyes, ears, mirrors, and common sense.
What’s really interesting is that in EVs, on longer trips, the slower-speed modes are actually associated with shorter overall trip times. You’ll spend a lot less time charging if you go 60mph than if you go 75mph, which totally outweighs the minimal time savings from driving 15mph faster.
Having taken my EV up past 130mph I know just how fast the battery drains while having fun. However, if having fun is the reason for the drive I’ll gladly wait extra time at a fast charger someplace.
Technology connections also pointed out you’re better off charging to only 80% even if it means more stops bc it takes so long for those last 20%.
This seems like it should be obvious, or maybe that’s just to me. I would offer that there might be a bit of helmet factor in here as well, where people tend to take more risks believing they’re safer, though I suppose that would relate more to the safety systems. I love that, in pondering a fix, they mention increased intervention when the solution is to just not have this BS. If driving is too hard, don’t do it. Oh, but “whaa, it’s so hard paying attention!” Except that you still have to with these nonsense aids. We as a society need to bring back the concept of personal responsibility.
Reminds me of the idea that the best thing we could do to decrease accidents is install a sharpened metal spike on everyone’s steering wheel. Obviously it’s meant to be absurd, but, also, maybe there’s some truth in that?
Speaking as someone who knowingly DD a car with a non collapsible steering column that was known to become a spear in a front end collision I can say such a spike wouldn’t help much. My opinion is also supported with the fact such spiked used to exist in the form of add of steering wheel assists and had to be banned (as did the non collapsible steering columns) because of the carnage they caused rather than fading out from lack of consumer interest.
I think there’s a balance somewhere between “too much” safety and complete disregard, but one of my major problems with humanity as a whole (not as much on an individual basis—I’m like the opposite of Larry David in that I generally like individuals, but hate humanity) is the tendency to swing from one extreme to another rather than finding and staying in a sensible equilibrium. IMO, we’re past the balance toward the “too much” safety side where they’re not only impinging on joy and freedom and have a high monetary cost in purchase and repair, but advanced CC, active electronic safety systems, and the increased weight and size of vehicles to handle passive safety seem to have gotten to the point where they’re working against their intent or creating a kind of unwinnable safety arms race. IMO, which carries no authority whatsoever, I think the best balance was before the high frontal offset crash test and implementation of these active driving and safety systems. (Except blind spot monitoring and rearview cameras, though we wouldn’t need those if we could actually see out of cars. A lot of that is a result of crash resiliency for side impact and rollover, but maybe cars with lower cgs could be allowed a reduced rollover strength and side impact wouldn’t have to be so difficult to excel in if other vehicles weren’t so damn heavy from so much safety engineering.)
I wonder what the trade off is between injuries prevented from massive pillars vs injuries created from not being able to see around massive pillars. I imagine the latter injuries involve more pedestrians, cyclists, and other small non-car things.
Me, too. Someone in a rollover with less protection due to thinner pillars would likely be at a lot lower risk of serious injury (provided they’re belted in so they aren’t thrown from the vehicle) than a pedestrian hit due to obstructed vision. Plus, the person in a rollover is at least partly to blame for their predicament, whether they’re driving stupidly or if they were hit by someone else and it wasn’t their fault, they still chose to drive such a vehicle. I buy small, low cars and still have obstructed view from pillars. Add tall cowls and hood lines typical of popular vehicles and it makes driving in an urban environment really demanding of attention and it’s still amazing how easily a person or cyclist can be hidden behind a pillar.
There was some rural intersection in the UK where the angles were just so that if you were driving at a certain speed and a cyclist was on the intersecting road riding a normal speed, the a pillar would hide them all the way to impact. If I recall, there were multiple fatalities. I know it was bad enough they redesigned the intersection
Brian,
Autopilot is hands-on. It’s the combination of adaptive cruise control and lane centering. It goes up to the speed it’s set at.
“Full Self-Driving” (Supervised) is the hands-off version. That software doesn’t have set maximum speeds and will speed in most profiles.
If the study used FSD, I’m not surprised. If it used Autopilot, I’m still not surprised. That has a driver selectable speed offset that most people set higher than the speed limit.
Musk isn’t one to be constrained by silly things like laws. The various adas in teslas are known to take the speed limit more like a suggestion.
Quick, someone do a comparison between Automatics and Manual transmissions so insurance can give me a discount because I’m paying so much attention (to not spilling my coffee in my hand while I shift gears)
This is a skill they should reward you for.
In the UK there’s a, not-entirely-unfair, stereotype that people driving automatics aren’t as good at driving. Mostly this is because getting an automatic-only license is seen as being easier than the full manual license, and most people know someone who had to get an auto license because they couldn’t pass a test while shifting gears.
Love the top shot. 71 Karmann Ghia Convertible going as fast as it could when I was 16. We don’t need no stinkin aids.
I have driven a Karmann Ghia over 90mph and DID NOT like the feeling of the body lifting off the road until the steering input went slack and I slowed down so I could steer again. NO thanks not ever again. I am good.
Not a top down convertible run, still stable at 110, the max I ever achieved alongside a Chevelle that had a speedo that went over 90.
The one I was in was a hard top very lifting body shaped.
I over reinforced the floor pan and added stainless alloy tubing to weak side rails, coil overs for rear suspension, Monza exhaust, and a new top. and that’s it. Off road, always landed flat, never bottomed out, and was still going last I heard with the 2nd after me owner.
The most important piece of data is missing from the article: how significant of a difference are we talking about? If it’s a highway average speed of 72.7 mph vs 72.9, it might be enough for someone at MIT to confirm their hypothesis and publish their findings but it hardly matters. I dunno — in my experience when the researches don’t give the figures in their write-ups, it’s because the difference is hardly significant.
~2mph faster in EVs on residential roads from week 2 onward (5-6mph over in EVs, vs 3-4mph over in ICE), and almost no difference on average speed on arterial and controlled access. (source: Figure 3 of the linked publication)