You are indeed conscious, and do indeed have thoughts. What the article discusses is that the overwhelming evidence seems to show your thoughts are generated by deterministic/proabilistic physical processes, and that in fact your conscious thoughts occur after your decisions are already made and often serve only as post-hoc justifications of what you already did.
Consciousness is just another component in the car. It doesn't determine or influence where the car is going. Perhaps the most valuable part of consciousness is it provides an effective way to use reflection which can communicate data about the internal state of the car to the other cars. (Often this data is objectively inaccurate, but still serves the interest of the car or the fleet.)
You do realize those aren't the same thing, right? In fact, they're basically opposites. Deterministic is "if/then", while probabilistic is "maybe". Something is either deterministic or probabilistic, and it matters which it is.
In your metaphor, consciousness is the learning system which is used to generate the model which drives the car. Our brains are semi-supervised learning, and our models get retrained when they drift far enough.
I say "deterministic/probabilistic" because I am thinking about physical laws; i.e., what we know of physics and quantum mechanics. The intent is to exclude the idea that consciousness is somehow above the rules of cause and effect. It probably would be safer to just say "deterministic" to prevent all the magic-shoehorning people sometimes attempt with QM, but I was being pedantic.
> In your metaphor, consciousness is the learning system which is used to generate the model which drives the car. Our brains are semi-supervised learning, and our models get retrained when they drift far enough.
No, in my metaphor, consciousness is totally unrelated to learning, as the article suggested. In my metaphor consciousness is basically just part of the network stack.
I recommend you read the works of the scientist upon which my pseudonym is based (not the Unicode part, that's just a dig at HN for the type of characters they allow)
Briefly, you cannot fully describe a complex system using merely the axioms of a less complex system (or more simplistically, all metaphors are (subtlety but importantly) wrong).
This is especially true of quantum mechanics, but in general it's a good rule to keep in mind. It prevents comfortable but naïve assumptions like "physics is just Newton because that's what we can perceive", or assuming that cognition is just summing neurons together.
In every aspect of life, you can safely assume that It's More Complicated Than That.
Although, after making that point, is there really a difference between a self driving car and a car and a driver in your example? Probably not. They're both being driven by something, and the answer shouldn't change based on what the driver is.