Uber crash update

_101730401_74f91528-cdee-4649-880e-3d90db136e03

We’ve got a bit more information on the Uber crash from a few months back. Firstly it would seem that the computer identified the victim as a cyclist (as she was pushing a bike) and thus it likely didn’t anticipate her stepping in front of the car. Why didn’t the “driver” then stop the car? Because she was watching TV at the time. So this kind of highlights a point I’ve long made about autonomous vehicles. A bored, inattentive driver will have mere milliseconds to react to a crisis, and they will almost certainly not be able to respond in time.

All of that said, the victim herself also isn’t entirely blameless. There were traces of drugs in her system, which probably explains her failure to look both ways before stepping onto the road. So it is starting to sound like a standard car accident, a car “driver” who wasn’t paying attention, a pedestrian who also wasn’t…..and an inexperienced learner driver (the computer) which misinterpreted the actions of a cyclist/pedestrian.

Indeed, this does highlight the problems a computer faces in making decisions. For example, take the two photographs below. I’m going to guess that anyone reading this would automatically slow down if you saw the little girl, but you might not slow down for the guy on the racing bike (in fact if you met him on an open road, you might speed up so you could overtake safely!). There is a level of judgement you’ll apply and its difficult to see how you teach that to a computer.

cyclists

Another example, I was coming down the A9 the other day (aka “the highway of death”) and I noticed a guy in an Audi coming down to a junction in front of me, guessing he won’t stop (like the big “stop” sign in front of him says he should!) I slowed down a wee bit and held my foot over the brake. And low and behold, he pulled out in front of me, forcing me to brake. Again, how do you teach a computer that certain drivers are cu*ts?

So there’s a long way to go before AI can fully take over from human drivers. I can see some possibility for using them, but probably only in a controlled environment, such as motorways, and where all the vehicles on the road are autonomous. Unfortunately, Uber’s main motivation seems to be simply at finding away of eliminating its drivers and saving itself some cash.

My guess is that the “driver”, while certainly at fault (thought its not entirely her fault) is going to get hung out to dry. Arizona is the sort of place where a hispanic on a low wage has about as much chance of getting a fair trial as you’d get in Salem in the 17th century. Its going to be all too easy to pin the blame on an obvious scapegoat and that’s what I suspect they’ll do.

Advertisements

About daryan12

Engineer, expertise: Energy, Sustainablity, Computer Aided Engineering, Renewables technology
This entry was posted in cars, cycling, future, robot car, technology, transport and tagged , , . Bookmark the permalink.

3 Responses to Uber crash update

  1. Warren says:

    With regards to your question, you already have the answer. Look for the (()()()()) Audi emblem and there you have it.

    • daryan12 says:

      Good point, suspect Audi might not be happy when they learn code is being entered into cars telling them to slam on the brakes when they see an audi, then take off his wing mirror as they overtake.

  2. Pingback: What “new tech” needs to learn from “old tech” | daryanenergyblog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.