Old myths about clean air needs and availabilities die hard, but here are some truths
Make it a habit to use only dry air (or gases) to inflate your tires. One can get away with less air quality on an automobile, but an intermittently used, heavy motorcoach pushes the limits of tire ratings much harder. From my experience, heavily loaded trailer tires are even more difficult to care for since the tire quality is lower and the tire design margins are even more challenged.
CO2 does not behave any differently in automobile and truck tires than dry air or Nitrogen does. The “CO2 leaking” myth comes from the butyl rubber used in many bicycle inner tubes. CO2 in bicycle inner tubes behaves differently because the CO2 is soluble in the Butyl rubber used to make the tubes durable in the bicycle application. CO2 permeates the full thickness of the Butyl rubber, swells the rubber and the CO2 somewhat rapidly diffuses through the rubber. So the little CO2 cartridges that bicyclists carry to rapidly inflate their road repairs or just harden up their tires, will last only a few days before the tire is low in pressure. The rubber in DOT tires is not CO2 permeable.
When initially filling with Nitrogen or CO2, much (most?) of the original atmospheric air is swept or displaced from the tire before the bead rim catches and the tire starts to pressurize. To be conservative, let's say that 50 % of the original atmospheric air (by volume) remains, however. Then from the gas laws, when the tire is fully inflated at 88 psig (6 atmospheres of pressure) less than one 12th of the volume (8%) may be comprised of original atmospheric air, with whatever moisture (relative humidity) the air actually contained that day.
If one takes a more realistic 10% of the original atmospheric air remaining when the bead catches and an endpoint of 100 psig (6.8 atmospheres), then less than one 68th of the volume (1.5%) is comprised of original atmospheric air. It's easy to see that a water-based bead lubricant “slathered” on (and into) a tire could easily contribute to more moisture being in a tire than the entrained moisture in the atmospheric air remaining in a tire if it has been initially filled with Nitrogen or CO2. I'm sure that's partly why Michelin advocates their water-less bead lubricant formula. Many of us use that for good reasons.
Most of the moisture in tires comes from “non-dryer” compressors concentrating the moisture entrained/existing in a standard atmosphere, as they compress the air to 6 or 8 atmospheres of pressure. In an industrial plant, the moisture coming out of a dryer (chiller) downstream of a compressor on a warm summer day is a steady flow of water. That's a different scale I know, but it will make you a believer in the moisture that needs to be extracted from compressed air.
BTW, Nitrogen is less costly to separate from air, but is more bulky to store (in gaseous form), thus it works for tire dealers but is less practical to carry on a vehicle. Liquid CO2 is more costly to produce but is highly convenient to store in liquid form in a small 10 or 15 lb. size cylinder on a vehicle. Thus the Power Tank popularity.