New Study Finds Siri More Distracting Than Google Now While Driving

[Neodym]

I've found Siri to be most useful for setting up appointments and reminders, and it works about 95% of the time. For everything else it gets distracting because it just doesn't get it and I find myself repeating my command.

I feel like there is also a difference in voice recognition quality between English and other languages. At least Siri improved there recently (with iOS 9 probably), so recognition of my native language today works far better than in the past, where it was more miss than hit. I can even mix it with English now (e.g. to have a song with an english title played) - that would never work during my former attempts. Overall it still feels a bit flakey, though.

 

[Lancer]

So the Mazda 6 is the most distracting.

I'm all for in-car entertainment but this study proves some modern cars have so much distractions today it's becoming dangerous to drive. Maybe we need smarter cars that take away the distractions allowing the driver to well drive.

 

[nope7308]

Me: Siri, play the Greenhorns debut album.
Siri: I can't find the contact Green Oranges.

It's a pure gimmick, through and through!

 

[cwt1nospam]

I spent 7 minutes in the car repeatedly trying to get Siri to call my mom last night.

I always have to ask for directions twice before it works. The first time has Siri cutting out mid sentence while responding to my request.

I was once told the way to get Siri engineers to actually listen to responses is to talk in a funny voice. What I need is a word for debugging that gets evaluated by the team. Short of that - I need a camera in my car to record how poorly Siri operates so I can post it on YouTube.

Yeah, if only your phone had a forward facing camera and a built in app to record the video... then we could all see/hear that you need to see a speech therapist.
🙄
Really people, the fact that the vast majority of people have no problem with it should tell you that maybe, just maybe, the problem is with the way you speak.

 

[andyyardley]

I had an Alarm called 'Wake Up', Sometimes i'd ask Siri to change my 'Wake Up' alarm to be say 6am, and i'd get the response 'Sorry I cannot find an alarm named wakeup'.

So I changed it to be called 'wakeup' and now she says 'Sorry I cannot find an alarm named wake up' 🙁

I've never had a piece of tech make me look and feel so stupid.

 

[andyyardley]

And while i'm ranting.

I don't understand why when I ask for something (clearly hands free as i'm using Siri) and she says here are some options the she doesn't read them out and she doesn't keep listening so I can say 'Option 2 please'. Instead I have to start hitting the screen, which kind of defeats the point of using Siri in the first place <crashes car>

As it currently stands it's the only major feature in iOS which really is rubbish.

 

[mantan]

So the Mazda 6 is the most distracting.

I'm all for in-car entertainment but this study proves some modern cars have so much distractions today it's becoming dangerous to drive. Maybe we need smarter cars that take away the distractions allowing the driver to well drive.

True. The smartest use of technology, though probably not feasible or economically viable at this point, is technology that would limit a driver's smartphone usage to specific voice activated technology. People trying to hold the phone to talk, read their text messages/emails, utilize apps etc while driving is the problem.

You can't have something heavy handed that kills all cell phone use -since the passengers should be able to do use their phones unabated.

But getting more integrated technology is half the battle. The other is forcing people to use it. Anybody who's observed other drivers in traffic know that at least 25% of them have their phone out while they are driving. Drivers have a mindset that they know they probably shouldn't, but nothing is going to happen to them if they are careful. The best safe driving technology in the world is useless if people don't use it.

 

[Kastellen]

Did they test whether any one of these is more distracting than two screaming kids in the back seat?

 

[deadshift]

From the study

To infer how distracting Google, Apple, or MS are, based on 65 drivers (potentially 20 for each phone system) seems way too small of a sample. Better studies needed here. Statistically you need a sample size closer to 400 to represent the US population.

No, you don't. Rough demonstrative numbers are achieved with this sample size. The margin of error should be shown, and it is significant, but the point is that these systems all suck, but there is a range. A more detailed study should only be needed when the systems stop sucking so hard.

 

[Kid Red]

Can I assume the same number of users for both platforms and their results were used? Otherwise, I would think more people using Siri vs Google Now could also lead to variances in results.

 

[dk001]

is this because siri can't understand u fully every time ?

Its better than yelling at your mate...

Siri can't argue with you 🙂 GPS does this ,,, the re-routung is good, but when i KNOW where i am enough to get back i still don't need some annoying little terd in the background saying "Your destination is on the right"....

ya i know where i live...it's my house after all ... Usually i just ignore the rest...

I figure if u can just block out what u *don't* want to hear anyway, then it doesn't really matter which is more accurate... however if u rely on it, then ya its just damn irritating at times.

Selective hearing? That is a married life skill (and some others too 😀 ).

 

[the8thark]

Sample size of 65!?! LOL. Talk about pointless, and laughable that they would even publish or announce their results. The most they should or could have said is that 'based on our study all of these voice-activated systems, all of them are distracting.' Also, this study doesn't account for 'Hey, Siri' or Carplay.

I agree. Pointlessly small sample sized survery. Anyone with a brain would just dismiss this waste of time survey.

 

[Kr15]

Looking at the statistics everything in these days are distractive to man at the wheel.

Also if product is rubbish you just don't use it and it becomes less distractive and vice versa.
So in theory we could look at this report other way round.

So in theory Google is less distracting just because it's less used?! 😉

I'm not pro Apple and against Google in general, don't get me wrong, Siri just fit's all my boxes and I had not the most pleasant user experience with Google.

 

[bcroft686]

I want to see the numbers for having a child in the car....

 

[AtheistP3ace]

I'm sorry did you say something? I was using siri.

 

[Makosuke]

I find asking siri for directions the most frustrating thing because it never gives me directions to the place I actually want to go on the first try. Sometimes I have to try twice, by the time the third time rolls around I just type it into the google maps app which never fails to understand what I want

There are obviously a lot of factors involved in good vs. bad directions and voice commands, but while I agree Apple has a way to go interpreting requests to specific addresses and guessing exactly what you meant in less-obvious cases, Siri has consistently done extremely well giving directions to specific chains or common types of store on request.

That is, if I ask "Give me directions to the nearest Dairy Queen" or "Take me to Costco", Siri almost uniformly executes that request correctly. Likewise, interestingly, it does pretty well with Yelp-answerable queries--if I say "How late is Sears open?", I get a correct answer for the local store. Basically, for the kind of directions or simple info I might need while driving, it does quite well, at least for me.

Asking about specific local businesses, particularly if they've got a very generic name or don't have a good Yelp listing, is a lot more hit-and-miss.

 

[dk001]

This was all nice. Does it really mean anything?
A side note, the NHTSA, States, and various insurance companies are not finding any statistical correlation with things like no Texting, and Hands Free laws leading to a reduction in accidents.

 

[Ethosik]

Why not just make cars one-seat? Having passengers is way more distracting than these. I always prefer to drive alone, leave the radio off, have my phone nowhere NEAR me. There is just too much stuff going on, and I keep my concentration 100%.

 

[dk001]

Asking about specific local businesses, particularly if they've got a very generic name or don't have a good Yelp listing, is a lot more hit-and-miss.

Unless Siri has really improved I can't say I agree. Last time I meaningfully used Siri for directions was when my wife was taken to the hospital via ambulance. Marina Hospital in Marina Del Rey California. Siri gave me directions to the Marina Animal Hospital near San Jose California. 5 hours away. The transit time of about 5 hours was the give away.

Haven't used it for anything critical or important since. I do play with it occasionally.

 

[ZubaZ]

Ouch on the Mazda 6. Shame, because it's a great car.

I do not see any shame in that. You have the power to choose not to use it. I have it, no accidents 🙂
IMHO everything is distraction while driving (even a nice commercial along the road).

 

[Makosuke]

Unless Siri has really improved I can't say I agree. Last time I meaningfully used Siri for directions was when my wife was taken to the hospital via ambulance. Marina Hospital in Marina Del Rey California. Siri gave me directions to the Marina Animal Hospital near San Jose California.

You seem to be agreeing with me. I was saying that asking for specific businesses--like the specific hospital you asked for--was hit-and-miss, particularly if it's not a chain, which is the same experience you had. "Where's Costco" seems to work fine; "Where is Industrial Electric Supply in San Francisco" may or may not come up with anything useful, particularly if the name is common.

I just did a quick test, for example, and got the following:

"Take me to [local hospital A]" gave several options which were basically correct, if a little confusing (multiple service listings for the same hospital).
"Where is [local hospital B]" gave completely useless results.
"Where is [local hospital B] in [City]" gave the correct result this time.
"Take me to the nearest hospital" accurately picked the closest one (hospital B above) and gave directions to it.

So the generic question worked okay, while specific ones were only right some of the time, depending on how I phrased it. This is in a rural California city, but I've gotten similar results everywhere I've tried it in the US while traveling.

 

[dk001]

You seem to be agreeing with me. ...

So the generic question worked okay, while specific ones were only right some of the time, depending on how I phrased it. This is in a rural California city, but I've gotten similar results everywhere I've tried it in the US while traveling.

I should have phrased the start a bit better.
In general I don't use Siri for local everyday places. Apple Maps has been getting much better detail in the last year for my area. Before that it was very hit/miss. Google Maps had the detail hence I used that.

After the hospital fiasco, I stayed strictly on GMaps.

 

[Frazzle]

My uncle Bob riding shotgun probably scores a lot higher on this scale than any digital assistant. So do my kids. Or a billboard with distracting content. Or eating a banana while driving.

 

[Bawstun]

I would suspect so, given how poorly and inconstant it works.

This.

 

[bsolar]

That is all well and good, I suppose. Except for one thing. It assumes a *linear* progression for distraction, frustration, etc. From the start, the test has introduced several additional sources of distraction, including things physically on the drivers body, that by themselves, I would find more distracting than answering a phone call, or the radio playing.

No, it doesn't. The reason of their study is exactly that it's not possible to simply evaluate a "progression of distraction" between different sources of distraction: it's clearly explained in the introduction of their paper, together with an explanation of the control experiments they did and how they validated their results.

Perhaps for the purpose of the tests in question, this doesn't skew the results? Hard to believe, but certainly possible.

It's only hard to believe if you either didn't read or didn't understand their study.