I was going to ask if you could plausibly get R, G, and B components that were all numerically greater from a sample object that physically reflects less light

I was going to ask if you could plausibly get R, G, and B components that were all numerically greater from a sample object that physically reflects less light of each color, using a real camera sensor and real firmware. It seemed to me that the comparison would at least be directionally correct, at least if the effect is in the same direction for every light frequency, even if it's meaningless in terms of any absolute scale. Then I remembered that a friend told me just today that cameras have gotten so smart about color correction that they may automatically correct an unusual scene (like a sky illuminated weirdly by wildfire smoke) to something more familiar and expected. So, maybe if a current smart camera thought that one of the objects was a specific kind of thing that "shouldn't" be a particular color, it could just totally replace the color with one it thought was more plausible!