β-FeSi₂, a semiconductor material made of two of the most earth-abundant elements, has important applications in thermoelectrics, photovoltaics and optoelectronics owing to its attractive properties such as suitable band gap and air stability over a wide temperature range. While point defects always play a vital role in semiconductor materials, only sporadic studies have been dedicated to the defects in β-FeSi₂. Here, using first-principles calculations we systematically investigate the intrinsic point defects in β-FeSi₂. Our results reveal that the formation energies of the intrinsic defects in β-FeSi₂ are high enough to prevent them from forming in a significant concentration under thermal equilibrium growth conditions. As a possible kinetic process generating intrinsic defects, we study the α-to-β phase transition of FeSi₂. We find that the phase transition is a slow process occurring on the time scale of an hour. Incomplete phase transition may lead to kinetically formed intrinsic defects. We further calculate the activation energies of the intrinsic defects and show that the experimentally observed conductivity of pure β-FeSi₂ should be a result of unintentional doping. Possible extrinsic impurities that may lead to n-type and p-type conductivity and their activation energies are calculated, which are in good agreement with available experiments. Our results provide guidance for optimizing the doping strategy of β-FeSi₂ for device applications.