The advent of genetically modified crops expressing insecticidal proteins of the bacterium Bacillus thuringiensis (Bt) revolutionized the management of insect pest species. Nonetheless, such disruption does not come free from controversies and concerns. The quick spread of Bt resistance in targeted pest species is an example, and the impact in non‐targeted arthropod species is another. The former is a continuous focus of attention, while the latter faces knowledge gaps and methodological shortcomings particularly when arthropod communities are concerned. If Bt crops such as cotton are aimed at drastically reducing a key community component, such as targeted caterpillars in the Neotropical region, a significant community‐level impact is expected and deserves assessment. However, the subject is neglected with existing studies focusing on arthropod assemblages or low‐level taxa compromising their resolution. Thus, food web analysis was used here to assess the short‐term impact of Bt cotton on the associated arthropod community in a Neotropical scenario. Bt cotton and related non‐Bt cotton were cultivated for two years, and the arthropods were sampled using sweep net and whole plant collections throughout the cotton phenological cycle. The gathered data were used to build (plant–consumer–predator) trivariate networks using food web analysis. Thirty‐six arthropod species were sampled with the prevalence of whiteflies, cotton aphids, boll weevils and green leafhoppers, the main pest species in the region. Cotton genotype and cultivation year significantly affected the abundance and diversity of the associated arthropod assemblage (Fapp ≥ 2.86, p < 0.01). However, the Bt cotton‐associated food webs were similar to those of the non‐Bt cotton, indicating no significant impact of this technology on arthropod food webs. This result would be welcome, if it not indicating that Bt cotton did not provide a significant control of the main target pest species of the crop raising questions on the usefulness of the technology in the area.