Assuming that you are dealing with a carburated engine... Fuel injected engines don't have venturis, of course.
The very idea of a venturi is to create a low pressure area in the intake stream. The higher the speed of the airflow, the lower the pressure will be. This works out well, since the higher the speed is, the more low pressure there is to suck fuel from the float bowl, so the venturi and the fuel jet control the air/fuel ratio quite nicely.
Yes, in carburated engines, particularly aircraft (Many of which are still carburated, as opposed to cars) which operate at high altitudes where the air is already very cold. Since the pressure drops, thermodynamics says that there must be a corresponding drop in temparature. In addition, the capacity of air to hold moisture is dependant on the temparature, so it becomes possible, if not likely that the air will fall below the dewpoint at the carb venturi. If the air is cool enough that this is below 0 degrees C, it forms ice, which stops the correct air/fuel ratio, and the engine doesn't run. Moisture isn't a problem though since as soon as the air passes the venturi, the pressure (and therefore temperature) goes right back up, and water droplets are reabsorbed.
This is indeed why you
sometimes apply heat to a carburator.
Carburators can't really adjust for air density, but they do adjust for barometric pressure, since it's a relative pressure drop which controls fuel.
Fuel injection systems need to have a better idea of how much air is passing through the system, in both volume and density. This can be done in several ways.
Some systems use an absolute pressure sensor. This determines the absolure pressure in the intake manifold, and by using the measured RPM and calculated volumetric efficiency of the engine, the computer is able to calculate the mass of air passing through the intake.
Most systems use an air flow meter. This is a device which measures the speed of incoming air through a metered chamber. These can use a variety of methods to measure the speed, and can signal the computer in various differnet ways, but their function is the same.
Either way, there will likely be other sensors for the air temperature at the point where it's being measured as well, in order to make gross corrections. This data is then used for a basic fuel calculation.
EFI systems then use an O2 sensor in the exhaust system. This sensor works by checking for hydrocarbons in the exhaust. (Not oxygen, which is kind of strange, considering the name.) The computer then uses this as feedback, to adjust the basic fuel calculation. This is also normally only done at part throttle "cruise" conditions. At wide open throttle, the computer will usually choose a deliberately rich mixture, since this produces more power, and reduces the danger of knock or detonation. This is why EFI systems don't need regular tune-ups like carburated engines do. They self adjust for changing component values.
Anyway, for cooling the system, using the A/C isn't a new idea, even the "waste" cooling from the A/C system. It's been discussed, shot down, tried anyway, and proven not to work at producing results.
For a turbocharged engine, getting rid of extra heat is a good way to make more power. (Not increase fuel economy, mind you.) When the turbocharger (Or supercharger, it's the same idea) compresses the intake charge, the air gets hotter. This is a natural consequence. Thus we use an intercooler to shed that excess heat back to the air, thus cooling the intake charge back to (ideally) ambient temperature, but at higher pressure. A common trick at a drag strip is to dump a bucket of dry ice onto the intercooler right before a run. The dry ice makes the intake charge very cold, so that the car can be set to make more boost without danger of detonation or knock. Again, more power, not better economy. People rarely care about fuel economy on the drag strip.

What was I talking about in the first place, now?