Near the stall, lift hardly depends at all on angle of attack, so the ailerons have very little impact on wingtip lift.
That doesn't follow. The ailerons don't work by changing the AoA of the wing. The aileron is a trailing edge flap that is deflected. If you look at the lift plots for such devices, you'll see that they are effective in increasing lift coefficient all the way to stall, but that the AoA (measured by reference to the original chord line) for maximum lift is marginally reduced.
There is a subtle difference between roll damping and roll wise inertia: The latter is resistance to rolling motion starting when a rolling torque is applied, but once a roll has started the roll wise inertia actually keeps the roll going. Roll damping on the other hand is resistance to the roll motion and will tend to stop the roll if it has started.
But what you don't mention, is that roll-damping depends on the positive slope of lift vs AoA. In the stalled regime, where lift decreases with AoA, this mode becomes an instability. And it's this that makes spin entry a significant risk if high roll rates develop at or beyond stall, probably even more significantly than yaw rates, although the two are of course linked.