Time-reversal is tricky, and a lot of logic for conserved quantities do not apply to it (for example just because the Hamiltonian respects time-reversal does not mean that there is a conserved "time-reversal" quantum number). There are hand wavy arguments, but the most general argument is through Kramers' Theorem (https://en.wikipedia.org/wiki/Kramers%27_theorem). This states that for a half-integer spin system with a Hamiltonian that is symmetric under time-reversal, every energy eigenstate has another eigenstate with the same energy related by time-reversal [1].
The addition of an external electric field to a system with electric dipole moment $\vec{d}$ will remove the degeneracy between the $+m$ and $-m$ states (which are related by time-reversal). Therefore the Hamiltonian describing a half-integer spin system which has an electric-dipole moment interaction cannot commute with the time-reversal operator.
Addition of an external $\vec{B}$-field will lift the degeneracy between the spin projections, and therefore the Hamiltonian for this half-integer spin system does not respect time-reversal invariance. This is to be expected though, as if we considered the full system under time-reversal the $\vec{B}$-field would flip sign as well. This is not true for an $\vec{E}$-field which would not change sign under time-reversal. Therefore an electric dipole moment in a half-integer spin system requires that the system itself breaks time-reversal symmetry.
This argument applies no matter how complicated the $\vec{E}$-field configuration is (so long as it's static). It also applies to gravitational dipole moments [2].
References:
- Sakurai, J.J., Napolitano, J. "Modern Quantum Mechanics", 2nd edition (2017)
- Weinberg, S. "The Quantum Theory of Fields, Volume 1 Foundations" (1995)