I'm not entirely sure what the answer to this question is. It's probably not friction heat alone as Parth Vader claims, since you can ignite coarse steel wool with the flame of a match, and yet the significantly finer "atomized" 100-mesh aluminum will not ignite under the same conditions ("atomized" refers to the manufacturing method of molten metal expulsion with inert gas through a turbulence-inducing orifice into a cooling chamber, with the word symbolic of the microscopic morphology of the resulting particles).
Similarly, anyone who has taken a blowtorch to a sheet of aluminum foil knows that it won't ignite, even though its surface area is much larger than steel wool.
The answer must have to do with some sort of chemistry, and in particular the rate at which oxide forms. Try the following experiment: mix equimolar ferrous sulfate and sodium oxalate solutions, filter the yellow precipitate, dry it, and then pyrolyze it in a test tube. The resulting iron powder bursts into flames in contact with any oxygen (although I don't have a reference for the particle diameter).
Meanwhile, a routine stroll through eBay or Firefox-FX reveals that aluminum powders with particle sizes on the order of 3000 mesh or smaller are readily available, and these are obviously not pyrophoric.
So while pyrophoricity is probably the main culprit, I have no explanation for why. Aluminum is far more electronegative than iron, with a value of 1.5 versus 1.8, and is well-known to have a much higher enthalpy of combustion (both on a molar and on a per-mass scale). The mean oxide film thicknesses of iron and aluminum are also supposedly comparable, on the order of 3 nanometers, so unless carbon steel acts differently, it can't be due to the oft-cited thin-film passivation property of aluminum.
However, I'm willing to bet that someone who works in materials science might have something more to add.