The research is aimed at human-artificial intelligence interaction capacity and threats. System analysis of threats arising from the exponential development of AI technology is presented. It is used scientific method of deducing essence along the way of tracing history of its origin followed by extraction of phenomena from essence. Analyses of AI technologies achievements, limitations, and forecasts is based on experience of creators. System analysis of 30 threats provided categories’ criteria including sphere, reason, probability, level, target group and impact; mitigation team, means, actions and efficiency. It is proposed innovative approach to overcoming risks based on the concept of human-artificial intelligence synergetic interaction embodied in technology, including the development of threats matrix (features and weights), profile (structures), continuum and virtual polygon of threats mitigation by Backpropagation method and technique of superimposing artificial and human sense-cognitive structures as activated neural networks for cocreating multi-actor/sector/disciplinary solutions.