We developed a new strategy for disaster risk reduction for gravitational slope failure: we propose validating on a case study a simple method for real-time early warning of gravity-driven failures that considers and exploits both the heterogeneity of natural media and characteristics of acoustic emissions attenuation. This method capitalizes on co-detection of elastic waves emanating from micro-cracks by a network of multiple and spatially distributed sensors. Event co-detection is considered to be surrogate for large event size with more frequent co-detected events marking imminence of catastrophic failure. In this study we apply this general method to a steep active rock glacier, a natural heterogeneous material sharing all relevant properties of gravitational slope failure, and demonstrate the potential of this simple strategy for real world cases, i.e., at slope scale. This new strategy being theoretically valid for all types of failures, it constitutes a first step towards the development of a new early warning system for gravitational slope failure.