QFT
In the literature: open
Does the flat-space limit of AdS/CFT define the full S-matrix
In plain words
Shrinking the curvature of anti-de Sitter space to zero should turn its exact boundary description into flat-space scattering. Whether this works beyond approximations is unknown.
Precise statement
Determine whether the flat-space limit $R_{\mathrm{AdS}} \to \infty$ at fixed string length of $N=4$ super-Yang-Mills correlators (via Mellin amplitudes, Penedones 2010) yields a well-defined nonperturbative $10\mathrm{D}$ type IIB S-matrix, including energies where black holes form.
What would settle it
A proof that the limit exists and yields a unitary S-matrix, or an obstruction at high energies.