As seen in the video, I was able to bypass the Download feature of the site and get the link to the file and download it directly from the source of the image. There is also no metadata attached to the image, so without proper credit and citing from the poster, it may be hard to determine if the image was generated by an AI.
Me knowing a bit of webdev myself, I'm not quite sure if there's a way to “fix” or to otherwise solve this issue, but hopefully there's a way to add some form of authentication before accessing the file directly, or maybe small, subtle watermarks can be added to the original file before being displayed.