That's interesting (though we're getting a little far afield). In general, the gravitational mass of a sphere acts as if it's concentrated at the center of a sphere. And the atmosphere is part of of the Earth's mass. So smelting the alumina should have no effect on the orbit of the Moon, for example.
But it's different if you're looking at the the effect of gravity at a point within a sphere. My recollection is the it's as if only material at your radius or closer to the center counts (so, for example gravity at the center of the Earth would be zero). Trying to figure out the net effect of adding gas to the atmosphere on the amount of gravity affecting all the other gas in the atmosphere is hurting my head.
So I think you're right that the smelting would tend to reduce the net gravity affecting the atmosphere (because it would replace some mass that fully affects the atmosphere with mass that only partially affects it). But the effect would be less than that from simply reducing the mass of the Earth by the mass of oxygen liberated.