There’s been a lot of apocalyptic rhetoric about how ChatGPT spells doom for History and the humanities. I understand—or at least, I did. Until I used ChatGPT. If History dies, it won’t be ChatGPT that killed it.
Short thread, #twitterstorians 1/7
Short thread, #twitterstorians 1/7
I’ve had positive experiences teaching with ChatGPT, so I wanted to see how high it could fly. I selfishly(?) asked it to do research my very niche area of research, drugs and the US Civil War. Honestly, I wanted to know if it could replace me. 2/7
With much dread, I asked ChatGPT to research and write me an essay about drug addiction and the Civil War. It was an epic failure. Here's what it produced ⬇️😬 docs.google.com/document/d/1Jx… 3/7
Bottom line--ChatGPT wrote an essay full of fake “facts” and made-up “sources” that did not exist!
It made up info and created fake journal article citations to back up the fake info. Every source cited by the AI was fake. Literally. Every. Single. “Source.” 4/7
It made up info and created fake journal article citations to back up the fake info. Every source cited by the AI was fake. Literally. Every. Single. “Source.” 4/7
It also couldn’t analyze these “facts” to answer the “so what?” question. All this is shown in the above Google doc. 5/7
Ultimately, ChatGPT could not even do the basic, bread and butter tasks that historians teach our students to do. If we continue to do our due diligence—for example, fact checking—I think we have much less to fear at this time than most (including me) assumed. 6/7
Again, if History as a discipline dies, it’ll be because the lack of jobs starved it to death. Not because AI made historians irrelevant. 7/7
• • •
Missing some Tweet in this thread? You can try to
force a refresh
