If you are conducting a two-level experiment and wish to perform a few runs at the center point to detect any curvature and obtain your experimental error, what are some of the negatives of using this approach?

You do not know which factor, or factors, contribute to the curvature if it exists

Both of these

By using the center points as your experimental error in lieu of a replicate, you will not be able to detect if a factor causes a variance shift