# An Uber self-driving car killed a pedestrian in Tempe, AZ **URL:** \ **Category:** Miscellaneous and Personal Stuff I Must Share\ **Created:** [March 19, 2018, 5:56pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836 "2018-03-19T17:56:17Z")\ **Posts on this page:** 20\ **Page:** 3
**Author:** ![T-Minus](https://avatars.discourse-cdn.com/v4/letter/t/c77e96/32.png) [@T-Minus](https://boards.straightdope.com/u/T-Minus)\ **Post date:** [March 20, 2018, 7:00am UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/41 "2018-03-20T07:00:11Z")
If this kind of thing keeps happening with self driving cars and they end up being pretty dangerous; I can see them banning them off the road. They were a good concept, but if they prove to be dangerous I wouldn’t mind them getting them off the roads until they fix them. ---
**Author:** ![kayaker](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/kayaker/32/441_2.png) [@kayaker](https://boards.straightdope.com/u/kayaker)\ **Post date:** [March 20, 2018, 11:49am UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/42 "2018-03-20T11:49:09Z")
It would be interesting if the deceased pedestrian turned out to be a vocal critic of autonomous vehicles… ---
**Author:** ![TroutMan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/troutman/32/6721_2.png) [@TroutMan](https://boards.straightdope.com/u/TroutMan)\ **Post date:** [March 20, 2018, 1:41pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/43 "2018-03-20T13:41:17Z")
> [@GreenWyvern](#): > > I made a guess at what happened, based on the facts available, and then waited for further facts to prove or disprove the guess. > > \*\*This is how all criminal investigation proceeds, and this is how all science works. \*\* Well, the “facts” available at the time were that the woman was walking her bike across the road, which you chose to disbelieve and go from there. A criminal investigation would wait to view the video before hypothesizing on the cause. It’s not a biggie; I like speculating as much the next person (and without it, this board would be pretty damn boring). I was just amused by your “let’s wait for more information” comment when you obviously weren’t. > [@kayaker](#): > > It would be interesting if the deceased pedestrian turned out to be a vocal critic of autonomous vehicles… See, _this_ is the kind of speculation that is useful. 🙂 ---
**Author:** ![Machine\_Elf](https://avatars.discourse-cdn.com/v4/letter/m/82dd89/32.png) [@Machine\_Elf](https://boards.straightdope.com/u/Machine_Elf)\ **Post date:** [March 20, 2018, 2:20pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/44 "2018-03-20T14:20:47Z")
> [@T-Minus](#): > > If this kind of thing keeps happening with self driving cars and they end up being pretty dangerous; I can see them banning them off the road. They were a good concept, but if they prove to be dangerous I wouldn’t mind them getting them off the roads until they fix them. [As of two years ago,](https://www.fastcompany.com/3055356/the-first-study-of-self-driving-car-crash-rates-suggests-they-are-safer) the track record for autonomous vehicles was already better than for human drivers. It’s surely improved since then, and will continue to improve. Funny, one autonomous vehicle hits one pedestrian on one day, and people start to think that AVs might be dangerous. But nobody seems to notice the [HUNDREDS of pedestrians across the US who are injured or killed by human drivers EVERY SINGLE day.](http://www.pedbikeinfo.org/data/factsheet_crash.cfm) This is like when the first Tesla Model S caught fire (and the driver calmly exited the smoldering vehicle); it got huge press, and few people mentioned the [hundreds of fires that happen every single day](https://www.nfpa.org/News-and-Research/Fire-statistics-and-reports/Fire-statistics/Vehicle-fires/Highway-vehicle-fires) in gasoline-fueled vehicles. ---
**Author:** ![scr4](https://avatars.discourse-cdn.com/v4/letter/s/59ef9b/32.png) [@scr4](https://boards.straightdope.com/u/scr4)\ **Post date:** [March 20, 2018, 2:26pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/45 "2018-03-20T14:26:48Z")
> [@PlaceboTarget](#): > > In my opinion, the threat artificial intelligent robots could pose in the future is not that they will become too smart to control but that they will be employed to handle situations they won’t be not smart enough to cope with. But we already employ humans to handle situations they aren’t always smart enough to cope with. How much does a machine need to be before they are allowed to do the same tasks? ---
**Author:** ![Spiderman](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/spiderman/32/230_2.png) [@Spiderman](https://boards.straightdope.com/u/Spiderman)\ **Post date:** [March 20, 2018, 2:37pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/46 "2018-03-20T14:37:39Z")
> [@aceplace57](#): > > It didn’t see the pedestrian. > > The public needs to realize autonomous cars will have lower accident rates. They won’t be perfect. Accidents will happen and people will get hurt or die. “Perfect is the enemy of good.” An investigation needs to occur, if it’s determined that the car should have seen the pedestrian & reacted, then these vehicles need to be banned until changes./updates render them safer. If the video shows the woman was standing on the curb for some time & then stepped into the street a millisecond before the car arrived there, then as good as the car is it can’t overcome the laws of physics to stop in time & that’s a different story. ---
**Author:** ![TroutMan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/troutman/32/6721_2.png) [@TroutMan](https://boards.straightdope.com/u/TroutMan)\ **Post date:** [March 20, 2018, 2:52pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/47 "2018-03-20T14:52:35Z")
> [@Machine\_Elf](#): > > [As of two years ago,](https://www.fastcompany.com/3055356/the-first-study-of-self-driving-car-crash-rates-suggests-they-are-safer) the track record for autonomous vehicles was already better than for human drivers. It’s surely improved since then, and will continue to improve. > > Funny, one autonomous vehicle hits one pedestrian on one day, and people start to think that AVs might be dangerous. But nobody seems to notice the [HUNDREDS of pedestrians across the US who are injured or killed by human drivers EVERY SINGLE day.](http://www.pedbikeinfo.org/data/factsheet_crash.cfm) This is like when the first Tesla Model S caught fire (and the driver calmly exited the smoldering vehicle); it got huge press, and few people mentioned the [hundreds of fires that happen every single day](https://www.nfpa.org/News-and-Research/Fire-statistics-and-reports/Fire-statistics/Vehicle-fires/Highway-vehicle-fires) in gasoline-fueled vehicles. It’s hard to make conclusive statements about their safety with respect to pedestrians at this point. Autonomous vehicles have logged fewer than 10 million miles on public roads, and there’s now one pedestrian fatality. The national pedestrian fatality rate is around 1 per every 480 million miles traveled. We obviously can’t draw conclusions from this tiny sample. AVs didn’t suddenly jump from infinitely safer than drivers last week to 50 times more dangerous this week. I agree they are probably already safer than drivers when it comes to avoiding accidents with other cars, but it might be premature to think they are safer in avoiding unexpected events like someone entering the road outside a crosswalk. We also consider AVs as one entity with a single safety factor, but it would be interesting to compare Uber to Waymo. Waymo strikes me as a bit more methodical in their testing and I wouldn’t be surprised if their cars handle some situations better than Uber’s. ---
**Author:** ![Morgenstern](https://avatars.discourse-cdn.com/v4/letter/m/f9ae1b/32.png) [@Morgenstern](https://boards.straightdope.com/u/Morgenstern)\ **Post date:** [March 20, 2018, 2:55pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/48 "2018-03-20T14:55:36Z")
> [@Machine\_Elf](#): > > [As of two years ago,](https://www.fastcompany.com/3055356/the-first-study-of-self-driving-car-crash-rates-suggests-they-are-safer) the track record for autonomous vehicles was already better than for human drivers. It’s surely improved since then, and will continue to improve. > > Funny, one autonomous vehicle hits one pedestrian on one day, and people start to think that AVs might be dangerous. But nobody seems to notice the [HUNDREDS of pedestrians across the US who are injured or killed by human drivers EVERY SINGLE day.](http://www.pedbikeinfo.org/data/factsheet_crash.cfm) This is like when the first Tesla Model S caught fire (and the driver calmly exited the smoldering vehicle); it got huge press, and few people mentioned the [hundreds of fires that happen every single day](https://www.nfpa.org/News-and-Research/Fire-statistics-and-reports/Fire-statistics/Vehicle-fires/Highway-vehicle-fires) in gasoline-fueled vehicles. You’re assuming it was an accident. But your evil neighbor has reprogrammed your self driving car to kill those unfriendly to our cause. **Maximum Overdrive**! It’s not too late to repent. ---
**Author:** ![Machine\_Elf](https://avatars.discourse-cdn.com/v4/letter/m/82dd89/32.png) [@Machine\_Elf](https://boards.straightdope.com/u/Machine_Elf)\ **Post date:** [March 20, 2018, 3:10pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/49 "2018-03-20T15:10:17Z")
> [@TroutMan](#): > > We obviously can’t draw conclusions from this tiny sample. AVs didn’t suddenly jump from infinitely safer than drivers last week to 50 times more dangerous this week. I agree they are probably already safer than drivers when it comes to avoiding accidents with other cars, but it might be premature to think they are safer in avoiding unexpected events like someone entering the road outside a crosswalk. “Unexpected” is exactly where AVs do better than people. [this AV](https://youtu.be/tiwVMrTLUWg?t=671) was smart enough to stop for a duck being chased by a woman in an electric wheelchair. Fast-forward to 11:42, and you’ll see another pedestrian incursion - this time a wayward toddler in a toy car pursued by his father, for whom the AV appropriately stopped short. If the car can see a thing, then it’s tracking that thing, making predictions about what that thing might do, and responding accordingly. Unless there was a hardware failure in the AV that hit the pedestrian early this week, it’s quite likely she popped into view at the last second, and was seen too late for the car to stop. If that’s the case, a human driver wouldn’t have done better. ---
**Author:** ![Jasmine](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/jasmine/32/2964_2.png) [@Jasmine](https://boards.straightdope.com/u/Jasmine)\ **Post date:** [March 20, 2018, 3:11pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/50 "2018-03-20T15:11:21Z")
I was disturbed to read that there is already legislation in the works to approve practical testing of autonomous cars WITHOUT a human backup driver. I think it’s a bit too soon to say the least. Yes, I agree with those who say that, if the machine makes an error, it’s unlikely that the human will notice, process, and react to it in time to prevent an accident. What if, however, the computer really malfunctions and the car begins to meander all over the place hitting people and things? If there is no human at the ready, do they have a kill switch that can shut it down and stop it and, even if they do, how much mayhem will ensue before the problem is recognized and dealt with? ---
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [March 20, 2018, 3:20pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/51 "2018-03-20T15:20:44Z")
> [@Machine\_Elf](#): > > Unless there was a hardware failure in the AV that hit the pedestrian early this week, it’s quite likely she popped into view at the last second, and was seen too late for the car to stop. If that’s the case, a human driver wouldn’t have done better. This is an assumption not supported by fact. And without any objective standard or verification testing, public safety is left to the whims of what a particular manufacturer considers to be ‘good enough’ in terms of accident avoidance. As it turns out, manufacurers tend to be focused on their bottom line and often view potential liability [in terms of the bottom line net cost](https://m.youtube.com/watch?v=SiB8GVMNJkE). As others have noted, the lack of fatal accidents on the part of autonomously piloted vehicles to date is a pretty meaningless statistic given the paucity of data and that for the most part autonomous vehicles have been restricted to driving at slow speeds in clear conditions with strong fail-safe measures in place to avoid any potential for an accident. A Google mapping car driving at 25 mph down residential streets in broad daylight which slows or stops when it detects pedestrians anywhere in its forward view field will avoid situations where an accident is likely, but isn’t very practical for a commuter needing to get to work on time. Stranger ---
**Author:** ![friedo](https://avatars.discourse-cdn.com/v4/letter/f/8edcca/32.png) [@friedo](https://boards.straightdope.com/u/friedo)\ **Post date:** [March 20, 2018, 3:33pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/52 "2018-03-20T15:33:56Z")
> [@Stranger\_On\_A\_Train](#): > > This is an assumption not supported by fact. Well, here’s some new facts. [Police chief: Uber self-driving car “likely” not at fault in fatal crash](https://arstechnica.com/cars/2018/03/police-chief-uber-self-driving-car-likely-not-at-fault-in-fatal-crash/) > [@Ars Technica](#): > > Herzberg was “pushing a bicycle laden with plastic shopping bags,” according to the Chronicle’s Carolyn Said, when she “abruptly walked from a center median into a lane of traffic.” > > After viewing video captured by the Uber vehicle, [police chief] Moir concluded that “it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway." ---
**Author:** ![TroutMan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/troutman/32/6721_2.png) [@TroutMan](https://boards.straightdope.com/u/TroutMan)\ **Post date:** [March 20, 2018, 4:06pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/53 "2018-03-20T16:06:37Z")
> [@Machine\_Elf](#): > > “Unexpected” is exactly where AVs do better than people… Yes and no. AVs can potentially see things better, especially when there’s a lot of data coming in. The example of the cyclist at the red light from the video you linked is a great example of that. But we can’t yet teach machines all the heuristics we use in making decisions. If the “unexpected” event is something we taught the machine, it will likely be better than humans at recognizing it and responding. Otherwise, a human might make a better decision based on experience. The duck in the road or toddler in a toy car examples aren’t impressive to me. Clearly there is an obstacle in the road, and no complex assessment is needed to decide if avoidance is needed. But what about a homeless woman (as it appears this woman was) at the side of the road? A driver might innately know that such a person is likely to be less aware of their surroundings and therefore more likely to suddenly enter the road, so the driver will show more caution. Is that knowledge taught to the AV? How can an AV be trained to recognize someone who is paying attention to their surroundings versus someone who is not? As said, we don’t have to get to perfect to be worthwhile. But it’s not as simple saying an AV will always react better than a driver. ---
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [March 20, 2018, 5:10pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/54 "2018-03-20T17:10:54Z")
> [@friedo](#): > > Well, here’s some new facts. > > [Police chief: Uber self-driving car “likely” not at fault in fatal crash](https://arstechnica.com/cars/2018/03/police-chief-uber-self-driving-car-likely-not-at-fault-in-fatal-crash/) When I see someone walking in the median, I slow down, check my opposite side for clearance, and if possible move over under the assumption that they might do something stupid or obtuse, particularly if it is a child or someone who may be impaired. Whether that would be sufficient in this case to prevent the accident we cannot say, but I’ll stand by the position that we need objective standards for verifying that automous driving systems meet at least a minimum criteria for safety performance and accident avoidance. Whether it is imposed from above or (preferably) developed by the people actually working in the field and adopted as an industry standard, without it there will be manufacturers who will try to cut corners to be early to market at the expense of lives and suffering. We have such standards in aviation, civil engineering, medicine, et cetera, and there is no reason to assume that ‘the market’ or precognition regarding product liability will cause most developers to use foresight and caution. Stranger ---
**Author:** ![PlaceboTarget](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/placebotarget/32/1073_2.png) [@PlaceboTarget](https://boards.straightdope.com/u/PlaceboTarget)\ **Post date:** [March 21, 2018, 6:54am UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/55 "2018-03-21T06:54:40Z")
> [@scr4](#): > > But we already employ humans to handle situations they aren’t always smart enough to cope with. How much does a machine need to be before they are allowed to do the same tasks? I think the comparison is misleading and abusive. Human beings are individuals with rights and obligations. A machine is a tool supposed to meet utility and safety criteria. The case of the self-driving car killing a pedestrian shows, in my opinion, that placing robots among the biggest potential threats to mankind due to their revolt capacity is quite silly. Devising inadequate machines and/or using them inappropriately seems a bigger threat. ---
**Author:** ![PastTense](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/pasttense/32/14550_2.png) [@PastTense](https://boards.straightdope.com/u/PastTense)\ **Post date:** [March 21, 2018, 7:39pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/56 "2018-03-21T19:39:14Z")
Here is an interesting possible explanation: > [@](#): > > But it’s one thing to recognize a human shape. Already they come in a variety of shapes. Size is one variable (adults, children), poses another one. Humans not always walk up straight, sometimes they are bent over, they sit, they crawl, and show many other crazy poses. Still, autonomous cars must recognize humans, no matter how bent they pose. > > And it doesn’t stop here. Humans often are packed with boxes or bags walking along roads. At Halloween many little humans run around in crazy costumes, making it difficult for cars to recognize what the heck tat thing in front of them is. Or they sit on bicycles, wheel chairs, Segways, or motor bikes. Each mobility device has a different movement dynamic. Motor bikes tend to pass between cars and are fast. Bicycles occasionally change directions very spontaneously… > > The victim of the fatal accident in Tempe was, so much is known, a homeless woman pushing a bike packed with plastic bags across a four lane street. The form of the object could have been confusing, it may not have had such a combination in its database of classified objects. > **[Was the Fatal Uber-Crash Result of a Classification Problem?](https://thelastdriverlicenseholder.com/2018/03/21/was-the-fatal-uber-accident-result-of-a-classification-problem/)** > > The exact reason for the first fatal crash with a self-driving car is still unclear, but the details that we know so far could indicate a problem area. The classification of objects. Autonomous car… ---
**Author:** ![Maserschmidt](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/maserschmidt/32/18829_2.png) [@Maserschmidt](https://boards.straightdope.com/u/Maserschmidt)\ **Post date:** [March 21, 2018, 8:30pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/57 "2018-03-21T20:30:36Z")
> [@PastTense](#): > > Here is an interesting possible explanation: > > [Was the Fatal Uber-Crash Result of a Classification Problem? – The Last Driver License Holder…](https://thelastdriverlicenseholder.com/2018/03/21/was-the-fatal-uber-accident-result-of-a-classification-problem/) Thaaaaat…is a very good possible explanation. ---
**Author:** ![mlees](https://avatars.discourse-cdn.com/v4/letter/m/ba8739/32.png) [@mlees](https://boards.straightdope.com/u/mlees)\ **Post date:** [March 21, 2018, 9:56pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/58 "2018-03-21T21:56:54Z")
> [@](#): > > The victim of the fatal accident in Tempe was, so much is known, a homeless woman pushing a bike packed with plastic bags across a four lane street. The form of the object could have been confusing, it may not have had such a combination in its database of classified objects. Why wouldn’t/shouldn’t the car slow or stop if it detects any object in the road? Why does it have to be an object it recognizes? ---
**Author:** ![TroutMan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/troutman/32/6721_2.png) [@TroutMan](https://boards.straightdope.com/u/TroutMan)\ **Post date:** [March 21, 2018, 10:14pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/59 "2018-03-21T22:14:08Z")
> [@mlees](#): > > Why wouldn’t/shouldn’t the car slow or stop if it detects any object in the road? Why does it have to be an object it recognizes? I would assume the concern is with an object at the side of the road, not already in the road. A person at the side might enter the road so more caution is needed. An immobile stack of bags should not cause the car to slow down. ---
**Author:** ![iamthewalrus\_3](https://avatars.discourse-cdn.com/v4/letter/i/258eb7/32.png) [@iamthewalrus\_3](https://boards.straightdope.com/u/iamthewalrus_3)\ **Post date:** [March 21, 2018, 10:31pm UTC](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836/60 "2018-03-21T22:31:47Z")
> [@Spiderman](#): > > “Perfect is the enemy of good.” > > An investigation needs to occur, if it’s determined that the car should have seen the pedestrian & reacted, then these vehicles need to be banned until changes./updates render them safer. There’s an interesting liability difference here. If a person driving a car hits and kills someone, they’re potentially criminally liable for that conduct. We don’t just ban them from driving until they become a better driver. We put them in a cage for some number of years. If an autonomous vehicle does the same, should anyone go to jail? I’m not for sure saying they should, but maybe? If you can be criminally negligent behind the wheel, can’t you also be criminally negligent while writing software or designing camera systems? [Previous page](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836.md?page=2) [Next page](https://boards.straightdope.com/t/an-uber-self-driving-car-killed-a-pedestrian-in-tempe-az/810836.md?page=4)