Apple CarPlay, Android Auto distract drivers more than pot, alcohol, says study
iamroadsmart.com
iamroadsmart.com
even voice can be distracting when you have to repeat variations of the same command five times and keep checking whether it actually works.
I hope this kind of thing doesn't get banned entirely. when it works, I find that having a good navigation system allows me to pay more attention on actually driving the car.
I think you answered your own question.
Because of the problems I have with Siri, I end up using the touchscreen much more than I'd prefer. My latest defensive habit is to queue up enough audio for the trip and not touch it again until I arrive.
I would definitely prefer hardware buttons.
I like CarPlay, and found in the Mercedes models that had the little control wheel in the center console it worked well, because you had a physical switch to interface with CarPlay, but had the flexibility of the screen. I'm no expert, but I drive a lot of different cars and that so far is my favorite way of interacting with CarPlay.
I'd still prefer a simple dashboard with physical controls though.
In my car, if I want to increase the fan speed, I need to tap a small target, wait while the system lags out, and then tap a new target on a popup. Meanwhile, I’m trying to focus on the road at (effectively) infinity. It’s distracting and unsafe.
In my old, very well designed 1991 car, I could push the button to do this. It had two major benefits: it never moved and I could feel it. Touchscreens can’t replicate the feel of a button, but they could at least have the decency to keep controls in the same place.
This may well be better than using the touchscreen, but it's still bad.
But at least in my car I only need to select fan speed in the first menu then scroll do actually change it. This in a model x. But I was really responding to the "100% touch screen" statement.
In contrast, IIRC a 2007 Prius could do this type of voice command with no obvious difficulty.
The only bug I've found so far was when I said "open driver door" and it displayed "opening trunk" and it slammed it into my garage door!
...if you are from English-speaking country or you know English.
Does Tesla understand every language in the world?
Or: how would my father change settings in Tesla? Would he need to learn one English phrase ("Tesla/Alexa/Siri/Whatever, open Google translate"), then speak in his native language, then somehow speak the translated text to Tesla's system?
Currently in his car, he does that by pressing a button. Done.
I can literally tell it "Turn on driver seat heaters to 3 bacons" and it will set the seat heater to the max setting (1-3).
https://youtu.be/mV8h2doorrc?t=58
https://www.tesmanian.com/blogs/tesmanian-blog/voice-command...
So no, I disagree that physical buttons are safer than an NLP enabled infotainment/dashboard display where I can turn on/off and adjust things using voice. Without taking my eyes off the road and hands off the wheel.
I think with the latest update I can also control the active suspension via voice which is the only thing I still use the display for while driving except for hitting cancel on the navigation once in a blue moon. Cancelling navigation might also be available via voice command now, I haven't checked.
In any case, voice and physical buttons (on the steering wheel) is much safer than touching or buttons below the screen the screen regardless. At least I feel much more in control personally when I can keep my eyes on the road and not glance away / feel my way to a spot on the screen.
I'm in the market for a car, and I want something new or new-ish, but I can't find any cars I like without touchscreens all over the place. I'm now looking at used cars from around 2012, before all the touchscreen nonsense became pervasive, but still with most of the modern conveniences I'm in the market for. I'm not in a rush so I might just wait and see still, but I said that a couple of years ago as well and the touchscreen craze has just gotten worse I feel like.
In what way is it cheaper to install seatbelts than not?
That seems bizarrely counter-intuitive as a more-or-less standardized low-distraction interface already exists for car radios.
Okay, sure, you add the touch screen so you can do GPS and stuff, but is it really that hard to wire up a pre-existing car radio knob and buttons to an Android device along with some software to map twisting the knob to adjusting the audio volume and so on?
It just baffles me how little apparent attention was given to this sort of thing. I wonder if these people are dogfooding.
That said, car drivers really should get in the habit of making playlists so they can set up music and maps before they start rolling.
I'm sure it can be distracting, but it is also easy to set up with minimal interaction.
Ultimately, it comes down to individual behavior.
The problem with all this is that it just depends. If I check a message after stopping at a red light? No real risk. If i hit "call" whilst on an empty straight motor way for a hands free chat? No real risk.
If I try and compose an email on a country lane, or find a specific artist to play, or set a reminder whilst driving through a town centre... quite a lot of risk.
My car still has cassettes, so I have probably taken more risk changing over those (where is that bloody mixtape I want) than the limited phone use I have done
If this were true cars wouldn’t have touchscreens built in.
As a rider, I love the lack of distraction and can focus on everything going on around me, an important skill for staying alive on a bike.
I put a Sony XAV-1000 in my '07 Honda Fit recently, it was the only Carplay unit I could find with a physical volume knob. You can hold the volume knob down to activate Siri, so I never have to actually use the touchscreen. It's been very nice for Apple Maps/Spotify, I'd recommend looking into it.
(Personally, I don't think I find CarPlay any more distracting than any other in-dash screen; I'm not looking down at the car's built-in navigation system any more than I'm looking at Apple Maps on the same screen. But, I also tend to listen to podcasts while driving, so once they're playing and the map destination is set I rarely have much impulse to fiddle with things.)
I have yet to find a car with a better UX (aside from road handling, which I find equally good, but that’s another topic).
Touchscreens should only be allowed in cars with self-driving capabilities, and probably only on the higher levels.
In some lists the scrolling wraps around. In some it doesn't. Sometimes it's a click on an item. Sometimes it's a shift to the side.
And the systems I've seen (in BMW's DriveNow) have the worst feature of them all. The main screen where you have your selection of things to do (Nav, Car options, Media, Radio etc.) is not a fixed list, it's an MRU list.
So, if let's say radio is the very last item on the right side of the main screen, and you chose it, the next time you go back to the main screen, it will be the first on the left side of the screen. Now use Nav, go back. Now Nav is the first item and radio is the second item.
It is my long standing theory that people who design car interfaces have never seen a car in their entire life much less have driven one.
(And just hovering/slightly touching the key will show the current setting in the display.)
Well, the knob is fine - the CarPlay apps aren't. You can find the physical control by touch, but for any non-trivial operation you still have to actively watch the display as you fiddle the dial to get to the action (on-screen control) you need.
Which is dangerously distractive on the road and kind of irritating when parked.
Right now, CarPlay interfaces with steering wheel remotes in the standard way where prev/next buttons change tracks and volume +/- changes the volume. What Apple could do is this: add a toggle somewhere in the OS that remaps these to cursor controls. This would, without any effort required on the part of auto manufacturers, bring something comparable to the Mercedes control system to the majority of CarPlay compatible vehicles on the road, all via an OS update.
Their approach is far better than the all glass, cost saving Tesla one.
It seems like a feature that no one would really care about either way that car manufacturers just seem to think people want. It may even be because its cheaper than to R&D a user friendly tactile system with enough features vs just throwing software at it.
The Mazda does have the little wheel (it's down near the gearshift) and in addition the touchscreen on the display screen is disabled when you are in motion (which I think is a great idea). I adapted to it within an hour (I was coming from our older 2007 Honda Civic which has none of these)
The Honda does not have the wheel and you can use the touchscreen any time (which I think is a bad idea)
So a good compromise for you might be a Mazda (I think all cars will come with screens especially now backup cameras are federally mandated)
Honestly, on 101 and 280 I see drivers texting all the time and they're driving right in the lane, no problems at all except being slow to react to what happens in front of them (which makes it obvious they're on the phone). I don't know how they do it. Even glancing down at my phone instead of the center console loses so much state.
Pulling over to make changes may not be the most convenient, but I think for me at least, it's safest.
I dont know how people can text and drive. I usually see cars swerving and driving way too slow into a high way and it always turns out they are texting up a storm when you look.
Give me a D pad, an OK and a Cancel button. And make a phone API that works exclusively by that and the sounds and chimes it outputs.
I just want every supported app to be in a top level list that I arrow through and OK. Then within any app the options are all navigable in simple ways. Overall functionality is reduced.
Voice control is a secondary system altogether.
Anecdotally, over the years I've been driving, a majority of potentially collision-causing mistakes I've made (and that I've noticed) have been when I've been cognitively distracted by navigation systems and was looking for a turn/exit or thinking ahead for a turn/exit.
I've adjusted to always prioritize real-time safety even if it comes at the cost of missing my turn and having to double back to get where I'm going. I believe that following nav directions in real-time comes at a safety cost, and am skeptical that many people both realize this and make the appropriate adjustments.
this probably varies depending on your spatial reasoning skills. my brother can take a quick look at a map and remember the connectivity of all the major roads and then navigate from memory. whenever I try this, I end up in a stressful loop of miss turn -> find place to pull over to look at map -> miss turn again. stress has a comparable impact to driving safety as distraction in my experience.
Now I have a quite large navigation screen in the mid console and those optional arrows next to my speedometer. It really took some time to get me to not look and the centre screen. It took even longer to get used to just listen to what the computer lady says.
I can't understand how this centre screen has become a safe standard. Especially in countries like Germany where there is no speed limit. It's a toy for passengers.
Now, I don't care so much. GPS can route me back pretty easily if I miss anyway.
Some people are however, still unwilling to be delayed no matter the danger.
It does still happen sometimes, though.
To me, this is where GPS shines. If I miss an exit, it isn't a big a deal because the GPS will tell me the best route based on my current circumstances.
How distracting is it to play a specific song on a radio from 1999? Or a specific FM radio station that isn't already programmed?
You can, if the car is made that way of course, but the software supports it. I would not buy a car knowing it's only a touchscreen. At least in my car, the radial controls the focus on Android Auto and press it down to select. The screen is not a touch screen.
There's no question that physical controls for the most common tasks in a vehicle (there are actually very few) take less attention.
Additionally, it seems like one great promise of software in a vehicle would be to augment and improve upon that fact by design. But it's terribly obvious that this has not been a design goal ever. The goal has been to take the familiar UX of your phone and transfer it to your car so that the media component is immediately familiar (this is nice), but all the while we could have been making the act of driving safer too...
This is a silly argument, given the number of arguments that I’ve seen drivers participate in.
The problem isn't OS itself, but applications. Android application developers are frequently implementing custom graphical components, that do not handle "presses" and "clicks" and instead try to directly interact with touchscreen.
Google have made a number of increasingly aggressive attempts to work around the issue — for example recent versions of Android try to infer purpose and interactivity of custom UI elements instead of expecting developer to declare it. But that does not help, when everyone is using a handful of fancy libraries, hardwired to require touchscreen.
Incidentally, the biggest offender is not a third-party component... It is RecyclerView — Google's official item list implementation, distributed as a separate library. It's support for DPAD navigation is so abysmal, that Google's own engineers (from different department, I presume), had to create Leanback framework largely from scratch instead of relying on official RecyclerView components.
That is not as bad because you can choose when to start playing a song, so that you do it when traffic is calm. If you're driving under the influence, that's affecting you the whole time.
It would be interesting to see how these results look when the driver is performing tasks on a non-touchscreen setup. I can easily change the radio station in my car using steering-wheel mounted buttons, but changing the AC/heat is quite difficult without looking (in my Ford CMAX) because of the uniformity/location of the buttons. I'd guess the results would be similar to using a touchscreen setup.
The idea that Android Auto or Carplay is worse than drunk driving is dangerous and disingenuous and complete misinformation.
That's what the title implies. It's not what the study found.
I’m willing to bet a significant number of people don’t wait till it’s safe to use the system.
> If you're driving under the influence, that's affecting you the whole time.
Could it be argued that driving under the influence is only affecting you when you have an at fault accident that you otherwise wouldn’t have had.
I’m willing to bet the overwhelming majority of driving under the influence events are ordinary and uneventful.
Title is just misleading, as is often the case with this kind of studies nowadays.
Driving drunk is far far far more distracting and deadly than using carplay. Every Uber and lyft driver out there are running mapping apps. Every recent car is shipping with these (my wife's sister listens to her playlist on drive to work with it and uses the map).
I've seen drunk drivers. Carplay has nothing on booze. In an overseas country an ex-pat wanted to drive me drunk back to where we were going to be working. Within 2 minutes I'd taken the keys and started driving myself because he was all over the road - we had to stop so he could throw up.
I did not really dig into it, but these devices might actually be strictly worse from a reaction standpoint. At least until the driver is loosing consciousness.
Critical thinking means carefully examining evidence, even if it challenges "common sense".
The article headline and narrative tries to generate a "carplay worse than drunk driving" narrative when this is totally unsupported.
First, the slow reaction time was measured while the subjects were told to carry out tasks on carplay. These tasks were very specific. Like use the BBC iPlayer app (ugh!) to play a specific radio station. Find and play summer by Calvin Harris, send text messages etc. AND the setting were such that you could not use voice for some of this.
OK - summer is in a ton of song titles, so finding summer by calvin hariss is going to take some typing. Making users use a THIRD party app on carplay is another whole issue.
So yes - if you are a total idiot who disables voice control and is trying to type out complicated things on carplay and send messages using text - you are definitely going to be temporarily distracted. They presented little evidence that while in use fatalities increase because of this risk.
My own observation is that folks stopped at stoplights are HIGHLY distracted by both their phones and infotainment.
I use voice control which works well, and I only do three tasks - message wife I'm heading home, get directions to home (driving time and detours)and play podcasts. I say this while waiting to pull out of a parking garage. In terms of actual road risk this in minimal.
For a general risk comparison it’s percentage of time people are drunk driving x drunk risk vs percentage of time they interact with these devices x device risk. So, figuring out had bad things are during use is a very important factor.
We might read reports indicating that X number of accidents were caused by distracted driving, but distracted driving is very hard to prove because it is easy to deny. The only statistics we have on it are either when a driver admits to not paying attention, or when there is overwhelming evidence for it, such as video.
I think we can easily assume that most at-fault drivers are not going to willfully admit to liability for causing an collision (especially if a party to the collision was hurt or killed), so it is safe to say that distracted driving is highly under-reported and underrated as a threat to road safety.
> Imagine testing how distracting texting was for someone who had never used an iPhone before the test.
I understand analogies aren't meant to be strict equivalencies, but there is a large difference in severity here. Failure to operate an iphone correctly probably wouldn't cause a life threatening situation. Failure to operate a car correctly is far more concerning.
Here in Australia European vehicles aren’t required to put the indicator stalk on the right hand side of the steering column.
So I’m constantly indicating my intention to turn with the windscreen wipers in both cars, and constantly retraining myself, then I swap cars again.
Adding fatigue and other distractions doesn’t help.
Also highly depends on the cars setup. I've had some rental cars where the screen was terrible and definitely was a distraction. Example being Audi and their button controls for carplay.
So sure, I can do that. But it'd be better if the pain was felt by Apple rather than co-workers.
As a side note, I'm not a big fan of LCD displays in cars. I prefer analog gauges/dials/knobs, with minimal illumination for night time. At night, the less light you have inside the car, the easier you can see outside. Plus your eyes will be more well adjusted.
My friend had an old Saab with the "Black Panel" button, which turned off all internal lights save the speedometer. If other readings became relevant (e.g., running low on gas), that gauge or light would illuminate. I drove a Peugeot 208 a few years ago with a similar function. I wish I could find a vehicle with something similar, but design is clearly moving in the opposite direction.
The counter argument is that while it is easy to let go of the phone when things get tricky. You can't sober up at will.
And the counter-argument to that is that things can get tricky faster than you would expect. So tricky that even the time it takes for you to react and put your phone down or drop it still wouldn't give you enough time to prevent an accident.
People need to stop paying attention to phones while driving, period.
They understood, later. They weren't necessarily happy but when you explain to people "do you want me alive or do you want me to talk to you", their self-interest seems to kick in, at least.
The best hands-free reaction time in this study was still more than twice as bad as the effect from alcohol.
You simply can't operate a motor vehicle and a computing/communications device at the same time without seriously affecting your competency at both. Humans aren't built for that kind of multitasking. I don't anticipate any new technology which might change that.
Pick a number to call from your contacts, or type it out, or even pick a contact and type out a whole message using one hand without as much as a glance at the screen.
When you’re drunk, you’re drunk every second of the journey. You’re drunk at every intersection, you’re drunk at every pedestrian crossing, you’re drunk every meter of motorway.
They’re not comparable.
When it's working well it's not distracting at all. When you run into bugs it's very distracting.
Change it here:
Settings>Touch ID & Passcode>Allow Access When Locked
I drove around Spain for 2 weeks and didn't see a single car accident of any kind. More manual cars (occupy your hands) and fewer digital distractions. I saw 3 accidents on my way back from Oakland airport to my house.
US drivers need to get off their damn phones. It's so dangerous.
We need advancements in integrating with computers already so that our brains can interface with them directly.
The problem is that they are either hard to learn to use, or really ugly. It's always really bugged me that Google Maps on a phone (or in Android Auto) has tiny buttons that are impossible to click for the most basic tasks.
It doesn't have to be this way.
I'm certainly not condoning the practice, but texting and driving wasn't a serious public hazard until everyone moved away from T9. T9 was a deterministic input method, and while sure maybe you would look at the message before sending, each individual letter required little attention.
Back before smart phones / GPS navigation / etc took off I did a two month cross country trip using a Garmin 60CSx (handheld GPS with hard buttons) as a live map. In car-cities, I held the unit in my shifting hand, and checking the map was effortless as it was always in the state I expected it to be.
Holding the touch screen in your hand, and thus having more predictable positioning, is a setup up from using a touch screen mounted on the car's interior, which is moving relative to you from road bumps etc. Even worse are OEM touchscreens that have poor response. I'd love for more studies to be done on this subject to create better data, but I also wonder how much it will be fought by car manufacturers that have gone all-in on touchscreens in the center console (which seems to be all of them).
Most screens are placed in the middle of the dashboard, when it should be close to the gauges or nearly in front of you.
Source https://www.motorauthority.com/news/1121372_why-mazda-is-pur...
I remember cars used to either implement that feature, or outright prevented you from even pairing a new Bluetooth device when your car was in Drive. But then all of a sudden it seemed like automotive companies stopped and touchscreens became acceptable. Does anyone know what changed?
The NHTSA guidelines are here: https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/distraction_...
It's become unnecessarily dangerous if I'm using GPS for directions because now I have to continuously press "decline" and "not now" so I can see my directions. This has happened multiple times in a minute. Very dangerous - merely ceasing to ask after the tenth decline would probably save lives.
The study is flawed in that it lacks a control group and is missing information about relevance. You say people don't see the road when you tell them to look at a screen. Wow, color me shocked. But if you're going to demand change, you need to be able to put that fact into context. Like, what are the measurements when asked to perform the same action in the built-in head unit? And is this a contrived example, or does it represent typical use.
The headline has a similar problem. We already know that driver reaction time is effectively infinitely high for distracted driving. If the driver doesn't see the obstacle then they aren't just slow to react, they don't react to it at all... because why would they react to something that they don't realize exists? So you're comparing that fact to delays caused my chemical impairment? How? A driver with phone integration will have no impairment at all if he's not looking at the display at precisely the time of the incident, but intoxication has a persistent effect.
Even something as mundane as turning on the fan requires several seconds of concentration because it's al. It's not like a regular car with buttons and knobs that you don't have to think about. You have to actively look for it as you drive which makes it very dangerous to use.
As for Apple Carplay, I have as well, and it feels like the designers don't actually use it for driving. Some decisions are absurd. For example, a lot of the buttons you need to press are on the opposite side of the display (at least for cars with drivers on the left side). It's totally dumb that they are on the opposite side that needs to be hunted for.
In terms of safety, big knobs and buttons are best.
I can pretty much toggle or set any media and climate settings(even open the glove box) after a software update last December.
Using Apple CarPlay or Android Auto to add higher-capability infotainment to an otherwise unassisted vehicle does seem obviously risky and the data from the linked study appear to confirm that. My own anecdotal experience is that I am very distracted by Apple CarPlay in rental cars (my own vehicle is too old to be compatible). This is especially because the user interface is slow, with too much emphasis on animation rather than immediacy, and with too many behaviors that require waiting, such as resetting the map back to full-size.
Edit: Found a video from 2014. https://www.youtube.com/watch?v=7tcawBX41VU
It's highly distracting when you have bad UX and bad programming. And good software quality can't be legislated, in my experience at least the more regulation there is the worse the SW gets, because it gets written by bureaucrats, not software minded people.
Was it the constant commercials that sold you? The fruit logo?
Anecdotally, I find CarPlay to not be distracting at all, mainly because I don’t fiddle with it while driving. I set up things ahead of time or I have a front seat passenger to help.
I miss tactile buttons and knobs that you could develop muscle memory around.
I don't understand the motivation behind these studies. Is there some financial gain to be made from this. Or perhaps I'm being too cynical.
I don't think anyone is making that argument.
The argument is that you shouldn't be using your phone via any interface while driving, because it isn't safe.
The second option would be able to distinguish if you are driving your car or riding in someone else’s car or on public transportation. Assuming that you’re not paired to your non primary vehicle.
1.3 million traffic fatalities yearly (worldwide).
Good to know if/when new tech distracts the drivers and adds to that 1.3 number?
Some car vendors have better implementations than others, but I frequently see things happen like the connection suddenly cuts or audio playback becomes choppy or sometimes the dash crashes entirely.
I suspect if it was less buggy it would be a lot less distracting, personally.
Second of all, it's horribly misleading to compare the degree to which using these systems impairs your driving to the degree to which being under the influence of alcohol or drugs does so. I have CarPlay, and while I fully recognize that paying attention to it to change things pulls my focus away from the road, I'm typically doing so for no longer than 2-5 seconds. Being drunk or high lasts much longer than that, and there's no way to just turn it off. Hell, even texting or a phone call takes orders of magnitude longer than that.
I don't know if there's a particularly scientific way to do this, but my gut feel is that a more useful number would be something like the integral of distractedness over time—so find the length of time of a typical phone call, and multiply the phone call amount of distractedness by that, and assign some reasonable length to the total car journey, and multiply the drunk/high distractedness by that. Then compare that to the CarPlay/Android Auto distractedness multiplied by 2-5s times some typical number of times people adjust it during a trip of the designated length.
Look down, interact, wait to see if you hit the right spot, try again, wait to see visual confirmation of successful input. Often all the above fleeting between road and panel.
Voice interaction is pure UI cancer in a number of ways and doesn't help
For instance, one stretch thru Dallas, my phone and CarPlay, which were connected, were giving me different turn-by-turn instructions.
Cliche: person with two watches doesn't know what time it is.
So I disabled CarPlay and just used aux-in for the audio from my phone.
the US navy is removing some screen-based controls after an accident -- this is less about distraction and more about the complexity of the interface, but still
https://www.theverge.com/2019/8/11/20800111/us-navy-uss-john...
mazda the car company made a similar move last year
https://www.motorauthority.com/news/1121372_why-mazda-is-pur...
Without comparable dosage this is meaningless.
Truth is driving requires one's full attention and any distraction is going to reduce safety, even in a self driving car.
I don't really find it remarkable that it's worse than pot, or alcohol within the legal limit though. They didn't offer much info on the weed thing and that surely depends on how "high" one gets, but if they kept it within the same range as the legal limit of alcohol (a small toke or two) that's no surprise at all.
I worked for years designing and building custom cars. Honestly, that giant screen in a Tesla is the craziest thing I've ever seen in a car, and I've helped build some pretty crazy cars.
Same here. I cannot understand how people are fine with it. I once had to drive a car who had it's GPS display further down (where the AC controls usually is, basically) so you had to glance down towards the shifting stick to be able to see where you are going. That feels like a similarly crazy idea. I'm lucky my car has a display that is as far up it can be without blocking the view out of the windshield.
Meanwhile, my car has physical buttons for everything and it took me about a week before I could control and navigate everything without looking. I could never do that with a touchscreen. I've had a smartphone since 2010 and still I cannot even type without looking, been practicing that for 10 years!
> It's impossible to learn where to press because there is no physical buttons to learn by.
You just pointed out glaring disadvantages with buttons/knobs. Situational Touchscreen Controls + Natural Language Processing voice commands is far superior. It's also more intuitive.
I can literally tell it "Turn on driver seat heaters to 3 bacons" and it will set the seat heater to the max setting (1-3).
https://youtu.be/mV8h2doorrc?t=58
If you're telling be voice commands (that actually work) are inferior when it comes to buttons, I don't know what to tell you.
When I was growing up my father was a "body man". He repaired wrecked cars. There was rarely a time when he had no work. He took me to his shop from the time I was born, so I grew up around crashed cars. I started working with him when I was 14 years old. By that time he was building custom cars in Hollywood, CA and was running George Barris's shop and I worked there and learned the trades and went on to run my own shop until I was in my 30s. By then I was building advanced driving systems for the severely handicapped. Mostly quadriplegics with very limited mobility who used power wheelchairs. We developed "Zero effort driving systems". Quite a few of my customers had been injured in automobile crashes.
All that is to say I have quite a bit of experience with people who've been in accidents. The fact is they can happen to any of us who drive or are in a car. But there are people who get into lots of them and it's accurate to say many accidents are the cause of distracted drivers, and that's often true even when it's not the driver's "fault". Accidents can be avoided by attentive drivers.
So the question to ask is "do touch screen displays increase driver distraction" and it's hard for me to imagine that huge screen in a Tesla does not. If a driver used it only once while driving that would have to qualify.
When we add to that the false sense of safety of a "self driving car" we further decrease the driver's tendency to pay full attention to the safe operation of a vehicle.
I worry about me making a mistake while driving, but not near as much as I worry about others on the road with me.
All that said, the Tesla is an amazing example of what tech can do in that arena and I admire what they've done. But I'd never put my or anyone else's safety in jeopardy by using that tech on public roads while I was in the driver's seat.