Assume the prop turns in the same direction as above. 

1) First Factor: Each blade, as it travels from the top downward, will sweep forward (relative to the engine). Equally, on its upward travel, will sweep backwards. In flight, the airspeed of the blades will differ as follows: As the blades travel downwards (and sweeps forward) on the left-side, their airspeed will be the Flying Airspeed, plus Rotation Velocity, plus forward sweep speed. 

On the right-side, as the blades travel upwards, their airspeed is Flying Airspeed, plus Rotation Velocity, minus rearward sweep speed. So, each blade has higher airspeed while on the left side (downwards), less on the right side (upwards). This induces more thrust on the Left Side, which tends to cause the motor to yaw to the right. 

2) Second Factor: As the motor is tilted over backwards at the top, the Angle-of-Attack (AOA) of the blades while on the left side, relative to the airflow, is increased (relative to a vertical prop disk), similarly the blade on the right-side decreases in AOA. This also causes more thrust on the left side of the prop disk, and less on the right, causing a right-yaw, adding to the first factor. 

3) Third Factor: Usually in a paramotor, the pilot's body shields some of the airflow to the prop. As the motor (and pilot) yaw to the right, the airflow comes increasingly from the left of the pilot's forward line, exposing more of the left side to cleaner air (less shielding on left) and causing more of the right-side to be shielded from clean air. This also causes increased thrust on the left side, and reduced thrust on the right side, further adding to the first two factors. The result, is a constant thrustline aimed to the right of the flightline. This causes the wing to bank to the left, even though the motor is trying to turn to the right. 

The difference between the two major effects, GP and ABT, is that the first is a momentary force, which disappears after the changing pitch, whereas the second is an almost constant force, with the pilot/motor combination almost continuously facing to the one side of the wing's flight direction. 

The first, GP, can suddenly appear sometimes unexpectedly, and can surprise the pilot in its extent and severity. The second, ABT, is almost constant, depending on the extent to which the prop disk is tilted relative to the vertical. 

ABT can lead to a continuos oscillation, with the wing rolling to the one side, until the pendular stability kicks in when you go "over-the-top", then rolling to the other side as your inertia takes you through the bottom of the "swing", then ABT pushes you back into the roll, which seems to get worse if the pilot attempts to do anything to counteract with the brakes. 

Both are quickly counteracted by reducing power. Read that line again! GP can be prevented by a healthy understanding of precession and preventing sudden changes in pitch. 

ABT can be prevented by a proper static hang-check, and setting the attachments points and/or Centre of Gravity to ensure the motor (and prop) is as close to vertical as possible. Remember that even a perfectly vertical static prop will tend to lean over backwards by as much as 15 degrees due to the thrust being 6 or 7 meters below the drag (length of lines separating motor/thrust from wing/drag). 

Thorough understanding of the three major propeller effects (Torque Effect, Precession and ABT) should be part of every powered paragliding training course. They account for the majority of incidents and accidents. 

I hope this helps prevents a few incidents; and assists the general knowledge, safety and enjoyment in this fledgling recreation aviation sport.