We conducted a survey following a modified methodology from that employed by Göring et al. (2023), reducing the survey questions to the realistic appeal of AI-generated species images. Similarly to the survey by Göring et al. (2023), participants were randomized from both the university community and the general public. The survey's primary objective was not to compare models but solely to assess the general public's perception of how realistic an image of an animal species could be. No subject variables were added, and no personally identifiable information was collected or stored, in compliance with Brazil’s General Data Protection Law-LGPD (Law No. 13.709/2018). The questionnaire was made available online and distributed among academic groups and the general public for 48 hours, receiving a total of 621 responses (see supplementary material S1 for the database). Participants were asked to rate the images as: 1 - Completely false, 2 - False, but convincing, 3 - Not sure if it is real or not, 4 - Real, but highly modified, and 5 - Real, only processed.