This has little to nothing to do with the specific company (although is a convenient deflection ATM). The very idea that pre-programmed computers can deal with driving is the fundamental problem. In time, people will realize that, but at the moment the "solution" is to "patch" and make more laws to hide the truth: AI has nothing to do with intelligence.
Ultimately they will blame humans, humans are the bugs in their code.
As far as I know all self-driving car programs use some form of machine learning, so they aren't exactly "pre-programmed" in the sense of good old-fashioned AI based on strict, pre-determined rules.
What exactly do you mean by "deal with driving"? Drive without a single accident, ever? That's obviously impossible in practice. Drive better than humans, who in 2016 killed 37,461 people in the US alone? I don't see how that would be impossible - human drivers have a limited field of view, slow response times (average time to break is ~2.3 seconds), and are frequently distracted, sleepy, drunk, etc.
Let me know when self driving cars can get sleepy, distracted or drunk. Until then, I'm much more concerned with "stop before hitting things 101". This "pre programmed cars must be better than X" is getting old fast.
This is one of those self correcting problems, and it's going to be fixed way faster than the startups pushing this "smart machines" propaganda are going to like.
I'm not the original poster, but maybe they just mean that as it stands right now, artificial intelligence refers to nothing that is actually intelligent in a way that an average person probably thinks about intelligence, and it's a little hard to see where the leap to actual intelligence is going to come. Maybe?
Ultimately they will blame humans, humans are the bugs in their code.