peer reviewed ; This paper focuses on potential errors when assessing the human toxicity of corn farming in Wallonia, Belgium. The USEtox method is applied to the farming of 1 hectare of corn. Local data are used for farming data and GaBi datasets for background data. Field emissions due to farming are calculated by the most prevailing models. The results in human toxicity, cancer effect, underline the large contribution of chromium (Cr) emissions. But when characterizing fertilizer composition, only the total chromium is measured and therefore unspecified chromium is used as emissions. However, it is known that chromium in the natural environment is mostly the non-toxic form Cr (III), which would greatly decrease the impact as the characterization factor for unspecified chromium is, in USEtox, the average of Cr (III) and the toxic form Cr (VI). The impact for human toxicity, non-cancer effect is mostly related to zinc emissions even if zinc is relatively harmless. The impact of pesticides is negligible in both cases. These results show that caution must be taken when examining/interpreting toxicity categories. © 2019 EDP Sciences.