For more than 60 years, the gold standard for assessing aversive conditioning in humans has been the skin conductance response (SCR), which arises from the activation of the peripheral nervous system. Although the SCR has been proven useful, it has some properties that impact the kinds of questions it can be used to answer. In particular, the SCR is slow, reaching a peak 4–5 s after stimulus onset, and it decreases in amplitude after a few trials (habituation). The present study asked whether the late positive potential (LPP) of the ERP waveform could be a useful complementary method for assessing aversive conditioning in humans. The SCR and LPP were measured in an aversive conditioning paradigm consisting of three blocks in which one color was paired with a loud noise (CS+) and other colors were not paired with the noise (CS−). Participants also reported the perceived likelihood of being exposed to the noise for each color. Both SCR and LPP were significantly larger on CS+ trials than on CS− trials. However, SCR decreased steeply after the first conditioning block, whereas LPP and self‐reports were stable over blocks. These results indicate that the LPP can be used to assess aversive conditioning and has several useful properties: (a) it is a direct response of the central nervous system, (b) it is fast, with an onset latency of 300 ms, (c) it does not habituate over time.