We know that $A = I_2 - MM^T$ and $M^TM = I_1$, which implies that $M$ is a unit vector. The matrix $A$ is a projection matrix, and for a projection matrix, the eigenvalues are either 0 or 1.
In this case, the eigenvalue $\lambda$ of $A$ can be either 0 or 1. Therefore, the sum of squares of all possible values of $\lambda$ is:
$0^2 + 1^2 = 1 + 1 = 2.$