This work discusses the application of techniques serving the purpose of prediction and detection of unfavorable man-machine interactions. Prediction criteria of Pilot Induced Oscillations (PIO) were applied during design process of experimental multi-modal fly-by-wire control system for small light aircraft. Estimated PIO susceptibility was further verified during real-time simulations and flight tests. The verification process included subjective pilot/operator expert opinions and the results obtained from the automatic detectors, used for the PIO identification in long sets of recorded data. The general idea of the detection algorithms is based on the Fast Fourier Transform (FFT). These are being presented in this paper in a form of detection applications along with the results of the flight experiments.