Hacker News new | ask | show | jobs
by userulluipeste 1865 days ago
When ADAS systems come into play, the primary concern is who's going to shoulder the blame when things fail. This autonomous driving thing is no simple nor incremental improvement, this is an overhaul in our current structure of liability. So far, with a human driving, when things go awry, blame can be easily assigned, it's simple and case can be closed. When we introduce total absolution of the former party in charge, simeone else has to step in. The only one with any control here is the producer of the vehicle, yet the scope of responsibility is far too large for anyone at the traditional levels of rigor. The consequence is that ever increasing levels of safety requirements get demanded from the self-driving technology, enough to make it at any time both insuficiently mature and an even more desired thing for how much it has to offer in order to cover the said requirements. Mr. Musk seem like a smart man to not be aware of this, which makes me think about the moral hazard of his tweets.
2 comments

I'm assuming you are talking about when we reach level 4+ right? Anything below that is clearly still in the realm of driver responsibility. Level 4 is really where the large grey area will live, because level 5 will certainly require moving that liability to the system manufacturer.

Just a minor nit, "ADAS systems" is redundant.

Liability is the difference between L2 and L5 systems. Most people don't understand that Tesla FSD will operate in L2 mode until it's 99.999999% reliable. Until then, you'll have to keep your hands on the wheel and accept that you are liable, even if the car drives autonomously.