Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MacRumors

macrumors bot
Original poster
Apr 12, 2001
68,543
39,398



A new study conducted by driving academy and UK independent road safety charity IAM RoadSmart suggests that using Apple CarPlay while driving is more dangerous than texting or holding a phone to converse.

carplayreactiontime1.jpg

Drivers reacted 36 percent more slowly when using the voice function of CarPlay, and 57 percent more slowly when using CarPlay's touch functionality. Comparatively, texting caused 35 percent slower response times, and using a hand-held phone caused 46 percent slower response times.

The study found that using CarPlay or Android Auto with touch control caused drivers to struggle with controlling the vehicle's position in the lane and keeping a consistent speed and headway to the vehicle in front. Some of the study's other findings:
  • Participants failed to react as often to a stimulus on the road ahead when engaging with either Android Auto or Apple CarPlay - with reaction times being more than 50 per cent slower
  • Reaction time to a stimulus on the road ahead was higher when selecting music through Spotify while using Android Auto and Apple CarPlay
  • The impact on reaction time when using touch control (rather than voice control) was worse than texting while driving
  • Use of either system via touch control caused drivers to take their eyes off the road for longer than NHTSA (National Highway Traffic Safety Administration) recommended guidelines
  • Participants underestimated by as much as 5 seconds the time they thought they spent looking away from the road when engaging with Android Auto and Apple CarPlay via touch control.
It's worth noting that this CarPlay test involved just 20 participants, with 20 other participants using Android Auto, which is a small test group and may not be representative of the usage experiences of all CarPlay owners.

It's also not clear how familiar the participants were with the CarPlay interface before beginning the test or whether they had used it prior to the testing, but IAM RoadSmart says there was a "comprehensive familiarization process."

carplayreactiontime2.jpg

The test consisted of asking participants to complete three drives on the same simulated test route: a control drive, a voice-enabled drive (interacting with CarPlay via voice commands) and a touch-enabled drive (using CarPlay's touch controls only).

The route included two music-related tasks accessing music on Spotify and BBC radio while following behind a car (2.4 miles), two navigation tasks to a restaurant or a petrol station in a simulation of erratic motorway traffic (5.6 miles), and a figure eight loop done while reading texts and making a call (two miles).

Each participant was asked to react by flashing their lights when a red bar appeared on the screen, which was done to measure reaction time to an external event. These red light flashes appeared at four times during the drive when the driver was engaged with CarPlay.

Driving performance was measured by reaction time to the red bar, behavior measures like speed, lane position, and headway, eye gaze behavior, and self-reported performance.

Given the results of the study, IAM RoadSmart is calling on industry and government to "openly test and approve" CarPlay and Android Auto to develop "consistent standards that genuinely help minimize driver distraction." The full distracted driving study conducted by IAM RoadSmart can be read on the company's website [PDF].

Article Link: Study Claims Using CarPlay While Driving is More Dangerous Than Texting
 
Since Siri cannot understand a simple request 80% of the time, I can understand how this could lead to a distracting environment. Attention is required to make sure the AI can understand your request, or transcribe a dictated message. There are also so many bugs in this system, it usually requires troubleshooting when using CarPlay.
 
I should read the study before offering an opinion, but I won't...
You will not convince me that using your voice to control something is more dangerous than looking down at your lap and texting.
 
Last edited:
Too many variables and too small of a sample size to draw any meaningful conclusions. Were the CarPlay participants already familiar with CarPlay or were they trying to figure it out as they went? Also, I would seriously LOVE to see how these slowed reaction times compare to carrying on a conversation with an in-vehicle passenger and dealing with children in the backseat.
 
Based on my experience, using Siri via CarPlay can be incredibly frustrating. That's especially the case when trying to listen to a radio station.

For example, if I say, "Listen to <call letters of radio station>", it only works about 30% of the time. Sometimes, it works on the first try, other times, it plays some artist that has two of the radio station's call letters in his name (JD). From time to time, it plays something totally off the wall and I have no idea how it connected the call letters of the radio station that I spoke to what it's playing.

Since I can't see the transcription of what Siri *thought* I said with CarPlay (like I can on iPhone or iPad), I'm not really sure how to compensate for the fact that it didn't do what I wanted. Do I need to annunciate a particular letter more clearly? Do I need to slow down when I'm speaking? I have no idea. Sometimes, I have to repeat the same command five or six times before it finally does what I wanted it to do. Am I distracted during that time because my level of frustration with Siri is growing each time it gets it wrong? Probably.
 
I should read the study before offering an opinion, but I won't...
You will not convince me that using your voice to control something is more than dangerous than looking down at your lap and texting.
Worse than that -- taking one hand off the wheel for a significant amount of time to hold or type on your phone while looking at it and also trying to drive!

Too bad they did not include in the test for comparison "chatty" drivers that talk non-stop while looking over at the person in the passenger seat, usually using hand gestures while talking. And, lets see..taking your eyes off the road to read billboards. There are a slew of things that distract. Carplay isn't any worse than any of them.
 
Though much is made of hands-free phone in cars, it's really still dangerous, because the person on the other end of the phone is not sensitive to things happening around you. He can't see the roadkill you're about to swerve around, or the big 18-wheeler you've got to pass in a narrow space, or that you're trying to merge into a lane going faster than yours. A conversant in the car can see those things and can modulate their end of the conversation. A phone conversant can't. So now you're handling two things: the actual hazard you're driving around, and the disconnected conversation.

It's a real challenge.
 
I should read the study before offering an opinion, but I won't...
You will not convince me that using your voice to control something is more than dangerous than looking down at your lap and texting.

Exactly. As lame as touch controls/keyboards are, trying to give orders or talk to someone that can not understand you IS a major issue.

1584729435220.png
 

Attachments

  • 1584729428851.png
    1584729428851.png
    1.7 KB · Views: 152
never had a single issue with siri in my150. she understands me 100% of the time. granted all i do is send texts and listen to them.
 
IAM Roadsmart needs to get out of the business of doing any studies. It is hard to find more flawed reporting than what they are spewing... And that includes during this coronavirus event.
 
  • Like
Reactions: haruhiko
Doesn’t surprise me in the least.

CarPlay is very much a mixed bag and Siri continues to be a disjointed mess.

CarPlay is basically our main/only interface in our Chevy Bolt, so I get *lots* of exposure to using it...

I’d give it about a 6/10
 
I'm not too surprised, given how un-driver-friendly Siri is:
Use navigation >> ask Siri to do something >> Siri just stays there without ever returning to navigation, requiring you to look at/interact with the phone just to return to navigation, during which you may miss your exit or whatever.
 
I also find these results hard to understand, at least as compared to texting. Texting requires eyes off the road, fumbling with fingers/thumbs, not to mention mental distraction. I wonder if this is a matter of small sample sizes generating unreliable results.
 
Theres so much that bothers me about that graph but my biggest issue is with putting "drink-drive limit". Especially next to using cannabis, it suggests driving under the influence of alcohol. This implies Android Auto and CarPlay are far more dangerous. In fact, what the author is actually comparing the act of drinking in general. Overall, it seems as if the author is intentionally wanting to mislead and draws conclusions from an extremely small data pool that is unlikely to be representative of the general public
 
This study took place in the UK, where drivers use their left hand to interact with CarPlay screens. That may be their non-dominant hand and an extra challenge for their brain. I wonder if the findings are different in countries where people use their right hand instead.
 
I'm not too surprised, given how un-driver-friendly Siri is:
Use navigation >> ask Siri to do something >> Siri just stays there without ever returning to navigation, requiring you to look at/interact with the phone just to return to navigation, during which you may miss your exit or whatever.
This has always driven me nuts.
 
When I use Siri in the car, I find myself cursing at Siri more than anything else. Siri is almost two generations behind Alexa at this point. It's actually easier (and probably safer) to just type my query in than to ask Siri anything, including directions.

This is why personnel decisions are so important. Apple chose the wrong guy, and he chose the wrong approach.

Hopefully they have an internal team trying to replace Siri...but Apple's probably looking at their metrics incorrectly. For people that use Siri daily it's probably great. For everyone else that doesn't use Siri it's junk...especially compared to the other assistants.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.