It's an approximation, but a really good one. The reason it's really good can be seen by just considering how absurdly weak gravity is. For an electron gravitational effects are about $10^{43}$ times weaker than electrostatic ones, for a proton it's about $10^{37}$ (protons are more massive than electrons!)
So just to detect the gravitational interactions of, say, electrons, you would need to be able to design an experiment which could somehow notice this $\sim 10^{-43}$ effect.
It turns out that it's really easy to do such experiments: you get very large numbers of electrons, protons and neutrons and construct large, electrostatically neutral, objects out of them like planets and stars and people. And suddenly all the electromagnetic stuff cancels out and you can measure gravity.
But doing this on the scale of individual electrons or protons is almost impossibly (I hesitate to say actually impossibly) hard, because gravity is so weak.
Of course the fact that it is so hard is one of the reasons why quantum gravity is so hard: if we could easily do particle experiments where gravitational effects were easy to detect, then it would be easy to test theories of quantum gravity.
