For the same reason aircraft have $V_{S}$ — you don't want to get into stalled condition, because while recovery is possible, it is difficult and far from certain.
Due to the gyroscopic effect, the rotor blades reach the lowest point 90° behind the point of minimum lift. As the helicopter accelerates, the lift would decrease on the retreating side, which would push the aft side of the rotor down, pitching the helicopter up, and reducing the speed again (that is, unlike hover the helicopter is stable in forward flight).
Except the pilot wants to accelerate, so they keep pushing the cyclic forward, which increases the angle of attack on the retreating side (90° ahead of the desired tilt to compensate for the gyroscopic effect), compensating the loss of lift due to the lower air stream velocity.
So the angle of attack is higher on the retreating side because of pilot input, not inherent property of the rotor.
Now when the helicopter reaches the point of stall, the tendency to pitch up will increase, and it won't be possible to compensate it with more forward cyclic, since the retreating side is not able to generate more lift. In this sense, the condition is actually self-correcting as the helicopter will start climbing and losing speed. However
- the accompanying vibrations from the repeated stalling and unstalling of the blades as they rotate will stress them significantly and might cause structural damage and
- the violent pitch up accompanied with some roll to the retreating side (which it is depends on type) might lead to the pilot losing spatial orientation and consequently control over the usually low flying vehicle.
So you still want to avoid getting into the retreating blade stall condition in the first place.