Weak gravitational lensing is a powerful tool in observational cosmology. Large surveys of cosmic shear aim to constrain the cosmological parameters such as the mass density parameter $\Omega_m$ or the effective equation of state parameter $w$. Upcoming surveys, in particular EUCLID, will measure such a massive amount of galaxy shapes to such precision that systematical physical effects need to be accounted for. Ellipticity correlations are not necessarily only due to gravitational lensing but can also be caused by an intrinsic alignments of those galaxies, which can be thought of as nuisance “noise" on top of the lensing “signal". I will present one alignment model which our group uses to predict the intrinsic alignment of elliptical (i.e. early type) galaxies and discuss its implications on future surveys. Furthermore, I will show how we can turn that noise into signal — by comparing the intrinsic alignment correlations of early type galaxies to the weak lensing signal we can directly probe for gravitational slip, thereby conducting a test on general relativity on the largest scales.