No, it sounds like you're wrong on the facts of what the model did. It was *literally* trained on academic papers, and *only* returned information already in those papers. The virology students are encouraged to directly find, read, and discuss papers such as these.