This question is about the rotation of macroscopic objects and looks at the magnetic vector of an electromagnetic wave.
As basis for comparison, we consider an induction motor. The stator induces a magnetic field in the rotor. The field in the rotor follows the magnetization of successive stator poles and causes its rotation.
In this question, we replace the stator field by a circularly polarized electromagnetic wave. We replace the rotor with a small iron sphere (ball-bearing or similar) in weightless in empty space. We consider the magnetic vector that appears to be rotating.
This should induce a current and then a magnetic field in the sphere. This field would tend to follow the rotation of the magnetic vector of the wave and put the ball in rotation. Any momentum not captured would continue to be later intercepted by other conducting objects or even never.
For example, a LW transmission at 16 kHz would trigger a rotational speed of 960 000 rpm, speed attained by some electric motors.
According the law of conservation of angular momentum, the transmitter should undergo an equal and opposite momentum change.
According to the principle of reversibility: If you turn off the transmitter, the rotating magnetized sphere should behave as a transmitter.
If all this were to be true, there should be existing applications such as stabilization of spacecraft or transponders without active components.
This does not seem to be the case.
Edit, following comment: There may be a SETI application in which an object on the scale of one gramme would need to become detectable over long periods. nature.com/SETI-detection-by-data-carrying-objects Thus it could spin up reacting to natural radio background over a timescale of 10^8 years. If superconducting, magnetized and rotating near 1 000 000 RPM, would it generate a significant radio signal detectable from short distances ?