Es gibt einen Blogpost von Eine Sammlung von unverminderter Haarspalterei zum Thema ChatGPT, der nebst den (uns vermutlich allen bekannten) Argumenten über den Wert von Aufsätzen und ChatGPTs Beschränkungen hinsichtlich Fachverständnis auch ein Beispiel liefert.

We can demonstrate this by asking ChatGPT an almost trivially easy research question, like, “What is the relationship between Edward Luttwak’s Grand Strategy of the Roman Empire and Benjamin Isaac’s The Limits of Empire?” And so we did:

Image: Screenshot of ChatGPT’s response, asserting that the books approach the Roman Empire from different angles, providing a summary for each book, acknowledging that they intersect in some ways, but concluding that they’re ultimately distinct books providing different perspectives.

If you know nothing about either book, this answer almost sounds useful (it isn’t) (Footnote: the description of The Limits of Empire in particular is poor and mostly misses the book’s core argument […]). Now this is a trivial research task; simply typing ‘the limits of empire review’ into Google and then clicking on the very first non-paywalled result ([a book review]) and reading the first paragraph makes almost immediately clear the correct answer is that Isaac’s book is an intentional and explicit rebuttal of Luttwak’s book […]. A human being who understands the words and what they mean could immediately answer the question, but ChatGPT which doesn’t, cannot

Studys, die mit KI schreiben, ohne vorher die Fakten selbst zu ermitteln und hinterher das Ergebnis zu prüfen, schießen sich und der Zukunft akademischer Qualität in den Fuß.