Even then the camera is still controlling exposure level - how it achieves that in terms of combination of aperture and shutter speed could be varied but the meter would still be deciding the amount of light used to make the exposure - unlike in manual mode where you could very easily choose to under or over expose by over riding the meter. The idea is to make the exercise easy by reducing user input.

Yes a black, grey or white object with significant shadow or highlight areas would mean the results would be skewed - even then if you explain that in your assessment it will demonstrate your understanding of how a meter simply measures light and does not have any intelligent understanding of what the subject actually is.