Anthropogenic activities like oil and gas production or geothermal field exploitation are known to induce small magnitude earthquakes (usually M<3.0),which generally constitute more a nuisance than a real danger and which can be effectively controlled by close monitoring1,2.This type of induced seismicity is often confused with another much more dangerous type of induced seismicity – large event triggering (M>5.5) on nearby active tectonic faults3,4 – which has been so far mostly associated with the impoundment of artificial water reservoirs.In this case,the anthropogenic activities provide the small – but decisive - contribution that triggers the occurrence of destructive (or catastrophic) earthquakes which would naturally happen,but only decades (or more) later.We analyze the physical mechanism of the anthropogenic triggering of large earthquakes on active faults.We study the problem relying on experimental phenomenology,i.e.5,6 a) that sizable earthquakes occur only on active tectonic faults,b) that crustal stress values are those measured in situ and,on active faults,comply to the values of the stress drop measured for real earthquakes c) that the static friction coefficients are those measured on faults (and not in the laboratory or in bulk rock),d) that the effective stress values for triggering earthquakes are those inferred for real earthquakes.The conditions for earthquake nucleation are derived as a time-dependent solution of the Tresca-Von Mises criterion,while the time evolution of the process is derived from the equations of poroelasticity.Consistent with the issue that can be triggered only faults for which a drained path exists with the triggering source7,we find that the activation of active tectonic faults requires fluid overpressures smaller than 0.1 MPa.These compare with ambient stresses of the order of 300 Mpa and with deviatoric stresses of the order of 1-10 Mpa,yielding that the crust is always in a state close to failure5.Hence,earthquakes can be triggered by very small fluid overpressures,with normal and lateral slip fault mechanisms requiring a feather less than compressive ones (figure 1).The time dependent solution yields that fluids propagate as slow pressure waves along patterns operating in a drained condition (figure 2).In conclusion,we find that injecting fluids in the subsoil with the pressures in the practice of oil and gas exploration or storage may trigger destructive earthquakes on active tectonic faults within a few tens of kilometers.This makes earthquakes occur earlier – and never later – by a potentially substantial amount of time,and since neither the existence of a drained path nor the degree of maturity of faults are known,a safe defense requires to avoid anthropogenic activities in the vicinity of active faults.Relying on close monitoring and fast shut-off protocol is illusory,since the diffusion of crustal fluids has inherent time delays up to several years and any reaction to the first “precursors” is bound to be too late.