The use of Artificial Intelligence (AI) methods in Precision Decomposition (PD) of indwelling and surface electromyographic (EMG) signals has led to the recent development of systems that can automatically resolve most instances of complex superposition among action potentials. The remaining errors have to be corrected by a user-interactive editing process. Typically, 25% to 50% of such errors involve action-potential aliasing, whereby the action potential of a motor unit is incorrectly identified in signal data that actually supports the action potential of another motor unit. To drastically reduce this class of errors, we have added a new aliasing-rejection mechanism in PD algorithms. Experimental results on real EMG signals show that aliasing-related errors of the Precision Decomposition technique are thereby reduced by 80% to 90%.