A reconfigurable and machine learning resilient on-chip cryptography for graphene-based devices can be configured to utilize inherent disorders associated with the carrier transport in grain boundary dominated graphene field effect transistors (GFETs). For instance, a method can be configured to model a GFET as one or more physically unclonable functions (PUFs). A GFET PUF can also be reconfigured in a way that does not involve any physical intervention and/or integration of additional hardware components. A GFET PUF can be designed to operate with ultra-low power and can be configured to be robust and reliable against variation in temperature and supply voltage in some embodiments.