The problems and benefits of autonomous cars... 2 replies

Please wait...

FileTrekker Über Admin

I'm spending a year dead for tax reasons.

269,372 XP

15th December 2002

87 Uploads

22,521 Posts

1,748 Threads

#1 2 years ago

I was having this discussion the other day and my thoughts on this are;

While I think it's impressive that we are creating driverless cars, I don't think that we are anywhere near being at the stage in our maturity with technology to be able to trust it. Primarily the problem isn't the principle or even necessarily the technology itself, but the human being element is the major issue.

I cite the case where several Google employees were given a prototype of a driverless car and were made to sign a waver that stated that if they stopped paying attention to the road or be unprepared to stop the vehicle should anything go wrong, they would immediately be dismissed. Google found every single employee got so bored or ended up trusting the technology so much they all failed to do this.

Now if it was a given all things work perfectly all of the time, that would be fine, but when I think about how many times other consumer electronics end up developing errors, bugs, and problems, even if it is only in a small percentage of all users, that is now suddenly a situation that has gone from loosing a little data or preventing someone from doing some work to essentially having the potential to kill somebody.

Perhaps this will be a reality one day but I think our relationship and approach to technology and the stability of the software we put out has to change before this can become mainstream.

Danny King | Community Manager |

MrFancypants Forum Admin

The Bad

217,016 XP

7th December 2003

0 Uploads

20,003 Posts

6 Threads

#2 2 years ago

Eventually we'll get there. I think for now it would be good not to mix AI and human controlled vehicles. If you make AI control obligatory for highways and human control for cities you will probably end up with a system that is more secure for everyone. A similar method is used for aircraft wtih pretty good results.

Nemmerle Forum Mod

Voice of joy and sunshine

298,365 XP

26th May 2003

0 Uploads

28,147 Posts

5 Threads

#3 2 years ago

I think in the long run, keeping it expensive, having a few thousand people using the stuff on the road and using the feedback from that to improve... that's probably the best way to do things. We get the data and the deaths are kept to a relatively small number. That will work out eventually, you don't have to be a great driver to be better than most people - at least in the UK. I can't think of a journey I've taken where I haven't seen people doing stuff that if anything at all goes wrong, you're looking at a quite serious collision. Road deaths are not uncommon, the only amazing thing is that they're not more so.

For more general use, I think what's going to be the best use is to automate away bothersome or dangerous tasks. Automatic braking when someone's going to run into the back of another car... I see no reason not to do that, it's going to save lives. Adaptive cruise control - again, by forcing a more reasonable separation distance, I expect to save lives. For most people I suspect that automated driving won't be suddenly buying an automatic car as much as it would be... another feature having an automatic mode... and at some point that's all the features, the last feature goes, and you've got an automatic car.