Summary: In de Sitter space, the current induced by an electric field in vacuum is known to feature certain peculiarities, such as infrared hyperconductivity for light bosons in weak electric fields. Moreover, negative conductivity has been claimed to occur for light bosons in moderate electric fields, and for fermions of any mass in electric fields below a certain threshold. Furthemore, in the limit of large mass and weak electric field, the current contains terms which are not exponentially suppressed, contrary to the semiclassical intuition. Here we explain these behaviors, showing that most of the reported negative conductivity is spurious. First, we show that the terms which are not exponentially suppressed follow precisely from the local Euler-Heisenberg Lagrangian (suitably generalized to curved space). Thus, such terms are unrelated to pair creation or to the transport of electric charge. Rather, they correspond to non-linearities of the electric field (responsible in particular for vacuum birefringence). The remaining contributions are exponentially suppressed and correspond to the creation of Schwinger pairs. Second, we argue that for light carriers the negative term in the regularized current does not correspond to a negative conductivity, but to the logarithmic running of the electric coupling constant, up to the high energy Hubble scale. We conclude that none of the above mentioned negative contributions can cause an instability such as the spontaneous growth of an electric field in de Sitter, at least within the weak coupling regime. Third, we provide a heuristic derivation of infrared hyperconductivity, which clarifies its possible role in magnetogenesis scenarios.