Ouch!!! Scathing report from Consumer Reports.

Ouch!!! Scathing report from Consumer Reports.


“This isn’t a convenience at all,” says CR’s Fisher. “Monitoring the system is much harder than just changing lanes yourself. Using the system is like monitoring a kid behind the wheel for the very first time. As any parent knows, it’s far more convenient and less stressful to simply drive yourself.”

lunde | May 22, 2019

Obviously written by someone who may have tried said feature only once. My wife has nothing praise for the feature.

EVRider | May 22, 2019

Here we go again, until CR publishes another positive price about Tesla, then that's all we'll be hearing about.

Been there, done that. :-)

stephenfootball | May 22, 2019

I agree with most of their article. In it's current form auto lane change is not very good, TACC and auto steer are great and I use them all the time.

rkalbiarEV | May 22, 2019

They are not necessarily wrong. Unfortunately, I agree with much of what they are saying. The system is not very good yet but it is still early and will get much better. I am happy to be along for the ride and you know what.... If you don't like it - Don't use it. Problem solved.

Rutrow 3 | May 22, 2019

CR makes valid points. I understand that AP is Beta. The only reason I use Navigate on AutoPilot is to provide data to make FSD possible in the future. As it is now, it does change lanes/merge like an a$$hole. I usually override with the accelerator in moderate traffic.

Lovethe3 | May 22, 2019

I think all of it is great (and will get better, of course) but it takes practice to be comfortable with. A shame if this claim was made without finishing the learning curve.

dmastro | May 22, 2019

Their experience tends to match mine.

SteveWin1 | May 22, 2019

Yeah, I agree with most of what they said as well. I'm a little disappointing with the lack of progress lately. Doesn't seem like they're moving forward quickly enough to get close to FSD by the end of this year.

What happened to enhanced summon? That would be extremely helpful when its raining and you forget your umbrella.

efuseakay | May 22, 2019

When I first got the car, I was using AP pretty often. It's just not that safe right now. Too many random things can happen on the road, and you'd be in for a wreck. I haven't used AP in quite awhile.

Sparky | May 22, 2019

I agree with CR. In my opinion it's always been a mistake to marry the progress on automation with the progress on electrification. It will get better over time but right now the issues with automation are distracting people from the obvious benefits of electric drive.

LostInTx | May 22, 2019

Agree with much of the article. Yesterday I drove from San Antonio to Houston and NoA drove me nuts as opposed to driving the car efficiently.

The urge to move out of the passing lane is what drives me crazy. Yes, there's a place for it but I can see the 18-wheel rigs in the right lane much further than the car can. Moving from the left to right lane, only to approach the rig, slow down, go through double nag authentication before finally moving back over to the passing land is aggravating.

Also, on that stretch of I-10, the right lane was "grooved" ahead of being repaved, which makes the ride rough and loud. No, for the love of Baby Jesus, I do NOT want to move over, especially when there's nobody within sight behind me.

NoA needs to have an option turning off the incessant desire to move out of the passing lane in Texas for no reason other than a programmer out of California thinks it's a good idea.

*disclaimer* I love the car.

EVRider | May 22, 2019

I actually agree with the article too, even if it's a bit harsh. In my experience with NoA, lane changes are sometimes proposed when they shouldn't be, and not proposed when they should be (usually to move out of the passing lane), so I've never disabled the lane change confirmation option.

The last time I used NoA, I was in the rightmost express lane which was separated from the regular travel lane with plastic poles, and NoA proposed a lane change that would have taken me across the pole barrier. While the mapping data might have been bad, the lane markers under the plastic poles were solid, not dashed, so the car shouldn't have proposed a lane change even if didn't detect the poles. Regular auto lane change doesn't let me change lanes across a sold lane marker.

Hammonddave | May 22, 2019

I agree with CR. I use AP all the time but never have it change lanes by itself without confirmation. I use the AP stalk or the turn signals to confirm/initiate the lane changing. I also turn off the “use Carpool lanes” option. Too dangerous right now.

Garyeop | May 22, 2019

In KC I don't use AP a lot because of construction. Trips to is a life saver. An 8 hour trip is greatly improved using AP.

Kary993 | May 22, 2019

I agree with the article as well, but what is really lacking in the article is what HW3 is capable of from a learning perspective. CR just tries to say navigate on autopilot with auto lane change is dangerous, yet no one is getting killed using that feature. And to bring up three deaths using autopilot seems rather short sighted when thousand of ICE drivers are dying every week......CR is like wall street, they evaluate on old criteria and don;t understand what they are seeing and experiencing.

sjm4660 | May 22, 2019

LostInTx >>>>> NoA needs to have an option turning off the incessant desire to move out of the passing lane in Texas for no reason other than a programmer out of California thinks it's a good idea.<<<<

It does! I have the overtaking turned off so I, not the car, decides when to change lanes for passing. It's on the 'Mad Max' slider, change it to 'Never'.

kmartyn | May 22, 2019

Now that there is no longer a price penalty for ordering FSD after delivery, it might make sense for new purchaser to pass for now.

Mike83 | May 22, 2019

You should not buy it if you believe CR conclusions. We love it for oru two Teslas and use it all the time and know its limitations. It is also getting updated monthly with better and smoother lane changes. I especially like it to find the exit for us and take us off. This CR report is opposite of some YouTube reviews(I like to see the car in action). But I have stopped reading the toaster reviewer for a several years now. They don't get it and some people shouldn't get it either.

johnw | May 22, 2019

Holy cow, where are the rabid fanboys? I assumed all the posts would be bashing CR. I agree with the article as well- the feature is really still in beta and only Tesla-educated folks should really be using it right now. Only bad part of the article is they should have mentioned regular autopilot as well- THAT is fantastic and I suspect far safer than humans because most of the cars I see lately are weaving all over the place while their drivers on are on their phones.

JAD | May 22, 2019

CR is correct in stating that it is not better than a human driver at this point but no one is claiming that it is. like with most of auto pilot I know when it will work and when it won't. I would never think of trying to rely on it solely when in LA traffic.

The other thing not mentioned is that the new computer will be 10 times faster which will make it 10 times better or more which will be a huge improvement. self-driving will not be running on the computers currently being used, so why base its success on these computers.

Mike83 | May 22, 2019

JAD Exactly. Strawman arguments are so backwards. Love the gimmick and use it 95% of the time esp. on our last 2000 mile trip. FUD is crazy.

johnw | May 22, 2019

It's not great yet, but also not a gimmick. Magic 8 Ball, I'm tagging out with you ;)

Joshan | May 22, 2019

A magazine writing an article slamming a product that is in BETA? really?.... Does someone need to explain to them wtf a beta is?

Mike83 | May 22, 2019

I was used gimmick sarcastically. It really is amazing. Just check out the YouTubes on it.

texxx | May 22, 2019

Fairly accurate article, but I wish they had mentioned where AP does work well, and that's in stop-and-go traffic. It's too flaky for me to use at speed, but when you're stuck in a miles-long traffic jam it's fantastic.

Kary993 | May 22, 2019

@Joshan - Agreed....maybe CR should go back and watch the last big press conference and see if they can understand what the very intelligent engineers were actually saying.......but the CR wouldn't understand anyway....

johnw | May 22, 2019

@mike83 I wasn't aiming my secret Magic8 weapon at you, I could tell that was sarcasm.

Joshan | May 22, 2019

I think the only troll I see here is right above me....

Joshan | May 22, 2019

oops, 2 above me :P

Magic 8 Ball | May 22, 2019

This is all you need to know about the author of the article.

"Despite my love for quirky, old European sedans like the Renault Medallion,.........."
-Keith Barry

It does not help that a TESLA owner would refer to it as a scathing review. I didn't really read a review at all it was a hit piece. The author must be on drugs since as several have pointed out TESLA never claimed NOA to be better than a human so his entire hit piece was a set up from the beginning.

dmastro | May 22, 2019

Agree with @JAD

I think most of us acknowledge that NoA is beta and will continue to get better. The point is that its current iteration makes driving more difficult and dangerous than with a human driver.

Regarding the software being beta, yes Tesla calls it beta and provides all of the disclaimers to mitigate liability. But in the same breath, they heavily market and espouse the abilities of NoA. In fact, the Model 3 website shows a video of FSD, highlights all the features of NoA, and does not have a beta disclaimer on the main marketing or ordering pages... although I'm sure it's somewhere in the manual.

For the record, I like(d) EAP. The TACC, lane keeping, and driver initiated lane changes are great features. And I look forward to the day that NoA and FSD are ready for prime time.

Mike83 | May 22, 2019

johnw Got it and happy that I am understood by some . Those trolls do disrupt intelligent conversation. Strange brew.

texxx | May 22, 2019

@WantMY - WTF are you talking about? You don't even own a Tesla so how could you possibly know how AP behaves in slow traffic? Maybe TACC in your mythical too-special-to-name EV sucks in traffic, but Tesla AP in slow traffic performs exceptionally well, and I'm a fairly harsh critic of AP for general driving so this is no fanboi rant.

apodbdrs | May 22, 2019

CR is so biased it isn't even funny! I trust the TESLA system more than 90% of the drivers out there. I have been using it for sometime now and really feel alot safer. There so many ignorant people out there writing about Electric Vehicles and Auto Pilot, that have no business writing due to their lack of experience on any of them. I find so many just speculating on what they think these new features do! Very unfortunate the media even lets them contribute.

Effopec | May 22, 2019

I think it works fairly well in light traffic, but still as mentioned tries too hard to move back to the right lane with nobody around. In high traffic, high speed driving it is really aggravating. It is too slow to change lanes - will turn on the signal then wait 5 seconds or so before starting to make a move. By then, the gap has closed (maybe because the car behind you in that lane sees your signal and doesn't want you to cut in) so it has to abort. When I'm in this type of traffic I'll leave AP on but opt out of NoAP. Similar with stop and go traffic. AP is great, but NoAP doesn't change lanes fast enough to work through traffic. It should be better than you as it can see the whole blind spot and knows the closing speed of a car in the adjacent lane, or can abort if a car in another lane goes to take that space at the same time, but it is definitely in Beta at this point.

When in AP I wish it would put the toggles for auto lane change and confirm on the navigate screen so you can easily kick them on and off depending on what kind of traffic you are in.

calvin940 | May 22, 2019

CR has always had an Agenda with Tesla. This article appears to be no different based on what I am hearing (I don't read their stuff as of a few years ago because of their bias).

An objective piece of writing identifies all aspects of the subject not just those things that need improvement.

CR continues its path of irrelevance.

JAD | May 22, 2019

Why didn't they just compare it to the better systems the big guys have... Maybe because they don't have anything even close, so this is clearly the best drivers aid but not fsd yet just like it says everywhere.

Rutrow 3 | May 22, 2019

My suggestions for improvements:

- Smoother adjustments. The lane changes should be more fluid. They currently kind of "click" into and out of the lane. When I do a manual lane change, I gently drift to the next lane, passengers would barely feel it. AP should be better able to anticipate slowing traffic ahead and do a little coasting before regen has to kick in to maintain following distance.

- Better use of the width of a lane. The center of the lane should be preferred only when there is traffic in both adjacent lanes. Otherwise AP should bias toward the side of the lane away from other traffic and obstructions. When I am passing traffic, large trucks especially, I give a wide berth on that side of the lane, still with plenty of room to spare on the other side. There is great benefit to providing a larger buffer zone for reaction time if vehicles begin to drift out of their lane. This "lane center" priority causes a problem when an access ramp dividing line runs out to merge to the right lane because AP tries to center between the right side of the merge lane and darts toward the shoulder. I'd prefer that it instead strive to maintain a constant distance from the left lane marking instead.

- A little faster reaction to cars leaving the lane ahead. Cars ahead that move into the deceleration lane on our divided highways cause AP to hard decelerate and continue to do so well after the car is completely out of the lane. I get nearly rear ended because trafffic behind anticipates that I'll resume the speed when the car is out of the way. This is when I always have to override AP with the accelerator.

Thehowie | May 22, 2019

For those of us that don't have the "FSD" or "EAP" package, the message seems to be that we should wait. Based on most of the above, you mostly seem to agree. It sounds to me like CR did a service to its readers. They can't review promises.

slingshot18 | May 22, 2019

It won't be good until it's perfect. And that's the paradox someone will have to figure out. Until it's perfect, I don't want it.

FISHEV | May 22, 2019

CR is spot on...I feel like a baby sitter with the Model 3 vs. driving the 2015 Subaru Legacy with Eyesight. Tesla's in the shop getting bling, driving the Subaru for a few days reminded me how easy, convenient and dependable the Eyesight system is vs. Teslas.

CR is not the only one, Insurance Institute for Highway Safety does not include Tesla in its top safety picks with autobraking, lane keeping, blind side indicators and cross track warning while Subaru's are rated tops in every category they are in. And now the Eyesight system which was $1.5k is now standard on Subaru's, likely the cost to Subaru is in the $500 per car range.

Tesla has confused the EV/sustainable transportation mission with autonomous driving mission. Tesla charges $6k for the Enhanced AP system, 20% of the base car's price. If that is related to actual costs then had Tesla simply purchased Subaru's top rated safety system it likely would have been making profits on a regular basis for years vs losing $6k per car for Enhanced Autopilot. And the Tesla's would be cheaper, making EV and sustainable transportation system more affordable.

jim | May 22, 2019

Is a human drive better than NoA? Yes if the driver isn't distracted. CRs argument in favor of humans being better than NoA is hypothetical at best because it assumes people are not distracted, I'd trust NoA over the real-world drivers here in SoCal who spend most of their time looking down at their crouches while fiddling with their phones and yelling at their kids.

The article all but ignores the significance of the new chip which is likely to be a game changer if those autonomy day videos are any indication and instead, makes this snarky comment: “Tesla is showing what not to do on the path toward self-driving cars" yea, as long as you ignore the data that shows Teslas operating on auto pilot are involved in significantly fewer accidents than cars not using Tesla auto pilot.

One last thought: The article completely ignores the multitude of videos that show how Tesla has some amazing collision avoidance abilities that are beyond human reaction times and have clearly avoided collisions that a human drive could not have avoided.

FISHEV | May 22, 2019

"Is a human drive better than NoA? Yes if the driver isn't distracted."

As we've seen with the repeated fatal accidents where Tesla can't see turning tractor trailers and decapitates the distracted driver, even a distracted driver has a better chance without Enhanced AP vs. with it.

The Tesla system is way squirrelly if you are used to other systems. Just as CR says, the Tesla is dangerously unpredictable.

Hopefully Tesla takes the accurate criticism to heart and fixes it.

Hal Fisher | May 22, 2019

Yeah, it’s buggy and near sighted as hell. Would love to be in their development department. This car has so much potential, and seems to be lagging in development.

mos6507 | May 22, 2019

"Holy cow, where are the rabid fanboys?"

The ratio of realists to fanbois in this thread is a sign that we are finally and truly entering the trough of disillusionment phase of the hype cycle.

Mike83 | May 22, 2019

Owners are sick of your fud. Keep trying.

EAPme | May 22, 2019

Lots of calm and measured comments coming from a lot of folks who have experience using NoA. I like NoA but find myself relying on it infrequently, which is limited to longer trips (NoA's bread and butter).

I'd challenge to CR to provide some perspective by comparing Tesla's implementation to it's competitors. Class-leading, but far from perfect. Thing is, I don't think perfect is attainable if autonomous driving is combined with other, human drivers.

I worry about other knuckleheads more than I do EAP or NoA.

FISHEV | May 22, 2019

"I'd challenge to CR to provide some perspective by comparing Tesla's implementation to it's competitors."

That others are better or worse is irrelevant to addressing the flaws in Tesla's system.

jim | May 22, 2019


There were 3,166 driving fatalities related to distracted driving in 2017 alone. Four people have died in Teslas that were using auto pilot at the time. The facts do not support your position that "a distracted driver has a better chance without Enhanced AP vs. with it".

NHSA stats on distracted driving deaths:
Telsa auto pilot fatalities:

FISHEV | May 22, 2019

"Four people have died in Teslas that were using auto pilot at the time. "

Zero people have died in a Subaru Legacy for TEN YEARS. That's also on the National Highway Traffic Safety web site

Again, covering up flaws in Tesla's system by saying someone is worse is pointless. Consumer's loves the Tesla's but they are catholic in pointing out the flaws even in cars they like.