Elaborating what John Day said:
Actually, the pound per mile is not a bad approximation. Your oversight is forgetting that the carbon in gasoline is added, by weight, to two parts oxygen from the atmosphere.
The molecular weight of Carbon is about 12 g/mole, and oxygen 16 g/mole. This puts the molecular weight of CO2 at 44 g/mole with the carbon part making up only 12/44 of the weight.
Gasoline is a complex mixture of hydrocarbons boiling between 120 and 400 degrees F, with chemical formulas between C6H14 and C12H26, but a good “average” compound is C8H18. So, with hydrogen having a molecular weight of 1 g/mole, the 8 pounds of gasoline will have 96/114 parts carbon, and 18/114 parts hydrogen. This results in 6.74 pounds of carbon per 8 pounds of gasoline.
And 6.74 pounds of carbon will burn with 18 pounds of oxygen to produce 24.7 pounds of CO2. With a mileage of 25 mpg (which is the average fleet mileage for the 2017 model year – older cars have worse) the “close to a pound of CO2 per mile driven” is a darn good estimate.
Of course this is more high-school chemistry than 6th grade chemistry.