Your point still stands but the noise here is effectively a higher frequency modulation of the base AC oscillation.
In other words, the base oscillation of 50 or 60 Hz is not stable, but instead varies a little above and a little below at higher rates.
The magnitude is significantly lower than the base mains frequency. So possibly not audible above any other noise in the recording, but present nonetheless.
These filters will take that into account since it never has exactly been 50/60Hz. It's just the expected mean.
Maybe it's present in subsequent harmonics (e.g., 100/150/200Hz for 50Hz) as a systematic deviation from additive white noise. But, this would be difficult in practice to reliably isolate these signals via some form of component analysis.
In other words, the base oscillation of 50 or 60 Hz is not stable, but instead varies a little above and a little below at higher rates.
The magnitude is significantly lower than the base mains frequency. So possibly not audible above any other noise in the recording, but present nonetheless.