I think he's abstracting a level: advertising and fake news are both instances of misleading information. Sure, there are differences: fake news tends to be completely fake; advertising in the West is usually true to the letter but incomplete and often implying that the product will bring you more than the seller can really promise. Any good salesperson knows you sell lifestyles not features. Hence the scantily clad women we used to see draped over the cars in the ads (though at least in the 21st century we seem to have got a bit more tasteful on that front).
Given sufficient information I would buy a car based mainly on internal dimensions, fuel economy - yes the manufacturers are known for honesty on that one aren't they - reliability and price.
Reliability is exactly the kind of missing information he's on about imo. (Startup idea there, somebody please steal it!). Even for used cars it's hard to measure directly as nobody collects enough stats to do so. Example, what proportion of VW 1.9TD engines have failed due to conrod piercing the head unit? Is the risk any higher for engines with codes starting BXE? How does the risk break down by mileage and by age? If VW know these things they certainly don't tell us, and plenty of BXE owners want to find out (do a search...)
For new cars it can't be directly measured as there aren't any instances with 300k on the clock to evaluate. You'd have to make a proxy model based on component quality. What we do in reality is assume that brands previously known for reliability will continue to be reliable, but it's not exactly unusual for companies to trade their reputation for a quick buck.
If consumers were rational this would all be highly relevant. If, as you say, we aren't rational as a group then that's because of the car industry's success in marketing.