1) Micro controller input is high impedance
2) R to GND is very large compared to R in series
What this circuit effectively does is to filter the high frequency components of the rising edge of the PWM pulse. To the micro this is in effect a delay at the sample point. That is if you look at the PWM LO-to-Hi Threshold, the micro sees that threshold with some delay.
You have to be careful if the micro input does NOT have hysteresis. It might see multiple transitions at the switching point.
Why are you doing this? Are you reducing the edge rate of the PWM signal because you have signal integrity issues?