We’re not talking about human drivers. Do I need to post a video of an automated car holding up entire streets because someone put a traffic cone on its hood? If it can’t even deal with that most basic scenario then it isn’t compatable even to humans in its ability to make good driving decisions.
If the excuse for your terrible AI that endangers people is “That’s ok because human drivers are dangerous sometimes too.” That means the technology is not ready. That means you’re willing to cut corners on safety and put people’s lives at risk just to put out tech for tech’s sake.
Safety doesn’t have to do with people being assholes and trying to find ways to break them. It’s mostly about stopping in time and paying attention. The numbers validate this.
We’re not talking about human drivers. Do I need to post a video of an automated car holding up entire streets because someone put a traffic cone on its hood? If it can’t even deal with that most basic scenario then it isn’t compatable even to humans in its ability to make good driving decisions.
If the excuse for your terrible AI that endangers people is “That’s ok because human drivers are dangerous sometimes too.” That means the technology is not ready. That means you’re willing to cut corners on safety and put people’s lives at risk just to put out tech for tech’s sake.
Safety doesn’t have to do with people being assholes and trying to find ways to break them. It’s mostly about stopping in time and paying attention. The numbers validate this.
The above is an annoyance, not safety.
If it blocks a fire hydrant or stops emergency vehicles it is 100% a safety problem.
And, again, if it can’t even deal with that simplest of obstacles why the hell do we need to wait until it runs people over to admit it’s not ready?