If you have a matter field $\psi$ which is coupled minimally to the curved background, and a field $\phi$ which is coupled non-minimally, objects made of $\psi$ will move on different free-fall trajectories than $\phi$. This is a direct violation of the weak equivalence principle.
But the fact that the field $\phi$ is coupled non-minimally also means that you can observe its behaviour and essentially measure the local value of $R$. This is a violation of the Einstein equivalence principle even if you do not have a reference field $\psi$.
Now for the question of $M$: The reason why $M$ is the Planck mass is directly dependent on the context you are considering! The paper you link talks about $\phi$ being the Higgs field undergoing electroweak symmetry breaking which drives the inflation. Thus, $\phi$ reaches an essentially constant value $\phi_0$ in the post-inflationary era.
That is, the gravitational part of your action reduces effectively to
$$S_\mathrm{grav} = \int \sqrt{-g} \frac{1}{2}(M^2 - \xi \phi_0^2)R \mathrm{d}^4 x$$
in the post-inflationary era.
However, we measure experimentally right here and right now in the post-inflationary era that the gravitational term in the action is, at least phenomenologically, $ R/(4\pi G_\mathrm{N})$ where $G_\mathrm{N}$ is Newton's gravitational constant. In Planck units this term is written as $ M^2_\mathrm{p} R/2$. This means that if we want to fit the respective term in the action you give in the postinflationary era, we must fulfill
$$M^2_\mathrm{p} = M^2 - \xi \phi_0^2$$
This means that if we impose the phenomenological constrain, $M$ will never be equal to $M_\mathrm{p}$ for a non-minimally coupled field with symmetry breaking. The magnitude of $\phi_0$ depends on the details of the symmetry breaking and then imposes the range of $\xi$ which do not take $M$ too far from $M_\mathrm{p}$. However, if $\xi=0$, the phenomenological constraint gives immediately $M=M_\mathrm{p}$.