Two positions are held, both by serious people, and the atlas marks the question debated rather than picking one.
The first holds that a highest power was there before contact. Anishinaabe people say so; the concept sits comfortably in the wider Algonquian world; and early records report it. On this reading, missionaries found an existing idea and used the existing word, which is the ordinary way translation works.
The second holds that the translation did more than report. On this reading manidoo named a quality of power distributed unevenly across many beings rather than a being at the top, the superlative gichi- was an available construction rather than a title in use, and the pressure of needing an equivalent for God hardened a diffuse notion into a singular figure with monotheist contours. The strongest form of the argument points at what Gichi-Manidoo lacks: no narrative cycle, no ceremonial specific to him, no place, which is what one would expect of a category promoted to a person.
The difficulty is evidential and probably permanent. The earliest written Anishinaabe religious vocabulary was recorded by missionaries, in dictionaries and catechisms compiled in order to preach. The instrument that would have to be used to test the second position is the instrument the second position accuses. Nothing in the archive is neutral on the point, and no new source is likely to appear.
The atlas therefore reports both camps, names the ground each stands on, and leaves the question where it is.