Natural seismicity produces earthquakes that result in significant loss of life and billions of dollars of damage each year. Moreover, induced earthquakes from subsurface fluid injection endanger industries, such as carbon capture and storage, and geothermal, that are vital in our transition to a greener future. In this talk, the number and size of earthquakes that result from a fluid injection project in the Horn River Basin in Canada are linked to laboratory measurements of frictional stability, providing the potential to forecast induced seismicity before injection even starts. Following from this, laboratory measurements relating to earthquake nucleation and rupture propagation are incorporated into large-scale finite fault models that can produce sequences of earthquakes and help us understand why we get different number of small to large earthquakes. The results suggest that large faults, such as the southern San Andreas Fault in California, that today have very few earthquakes on them, potentially only rupture in large earthquake events and that the stress on these faults is very low.