The space physicists and the earthquake (EQ) prediction community exploit the same instruments – magnetometers, but for different tasks: space physicists try to comprehend the global electrodynamics of near-Earth space on various time scales, whereas the seismic community develops electromagnetic methods of short-term EQ prediction. The lack of deep collaboration between those communities may result sometimes in erroneous conclusions. In this critical review, we demonstrate some incorrect results caused by a neglect of specifics of geomagnetic field evolution during space weather activation. The considered examples comprise: Magnetic storms as a trigger of EQs; ULF waves as a global EQ precursor; Geomagnetic impulses before seismic shocks; Long-period geomagnetic disturbances generated by strong EQs; Discrimination of underground ULF sources by amplitude-phase gradients; Depression of ULF power as a short-term EQ precursor; and Detection of seimogenic emissions by satellites. To verify the reliability of the above widely disseminated results data from available arrays of fluxgate and search-coil magnetometers have been re-analyzed. In all considered events, the “anomalous” geomagnetic field behavior can be explained by global geomagnetic activity, and it is apparently not associated with seismic activity. This critical review does not claim that ULF electromagnetic field cannot be used as a sensitive indicator of the EQ preparation processes, but we suggest that both communities must cooperate their studies more tightly using data exchange, combined usage of magnetometer networks, organization of CDAW for unique events, etc.