Update.
-
Update. How often might this fear come true? A new study "found that [an] #AI text detector erroneously identified up to 8% of the known real [human-written] abstracts as AI-generated text."
https://www.sciencedirect.com/science/article/pii/S2153353923001566 -
Update. How often might this fear come true? A new study "found that [an] #AI text detector erroneously identified up to 8% of the known real [human-written] abstracts as AI-generated text."
https://www.sciencedirect.com/science/article/pii/S2153353923001566Update. Here's another study showing that tools to detect #AI-written text are easy to fool with "simple techniques to manipulate the AI generated content." But this one goes a step further and makes the right recommendation for teachers and schools. "GenAI tools and detectors…cannot currently be recommended for determining academic integrity violations due to accuracy limitations and the potential for false accusation."
https://educationaltechnologyjournal.springeropen.com/articles/10.1186/s41239-024-00487-w -
Update. Here's another study showing that tools to detect #AI-written text are easy to fool with "simple techniques to manipulate the AI generated content." But this one goes a step further and makes the right recommendation for teachers and schools. "GenAI tools and detectors…cannot currently be recommended for determining academic integrity violations due to accuracy limitations and the potential for false accusation."
https://educationaltechnologyjournal.springeropen.com/articles/10.1186/s41239-024-00487-wUpdate. More evidence that this fear has come true.
https://www.bloomberg.com/news/features/2024-10-18/do-ai-detectors-work-students-face-false-cheating-accusations"Even…a small error rate can quickly add up, given the vast number of student assignments each year, with potentially devastating consequences for students who are falsely flagged."
-
Update. More evidence that this fear has come true.
https://www.bloomberg.com/news/features/2024-10-18/do-ai-detectors-work-students-face-false-cheating-accusations"Even…a small error rate can quickly add up, given the vast number of student assignments each year, with potentially devastating consequences for students who are falsely flagged."
Update. This fear has come true to such an extent that students who write well without #AI assistance now feel pressure to use AI to "humanize" their writing and avoid the charge of AI-assisted #plagiarism. #Grrr.
https://www.nbcnews.com/tech/internet/college-students-ai-cheating-detectors-humanizers-rcna253878 -
Update. This fear has come true to such an extent that students who write well without #AI assistance now feel pressure to use AI to "humanize" their writing and avoid the charge of AI-assisted #plagiarism. #Grrr.
https://www.nbcnews.com/tech/internet/college-students-ai-cheating-detectors-humanizers-rcna253878@petersuber
This is all very unfortunate. It feels like the only reliable way to assess student’s ability to write a good essay is to put them all in a room with no tech other than pen and paper (i.e. an exam). This however misses testing other skills such as their ability to research a topic, and puts them under more extreme time pressure. Thoughts? -
@petersuber
This is all very unfortunate. It feels like the only reliable way to assess student’s ability to write a good essay is to put them all in a room with no tech other than pen and paper (i.e. an exam). This however misses testing other skills such as their ability to research a topic, and puts them under more extreme time pressure. Thoughts?@mattjhayes
I agree. Very unfortunate. I've retired from teaching and would not want to think up new ways to teach all that we used to teach through writing. I do see signs of creative thinking on that front. But I can't tell how effective it is. Some of it looks more like damage control than teaching. Some of it takes the form of suggesting, "use AI this particular way" (e.g. ask it to raise objections to your own positions and arguments, then answer those objections). Some suggestions along those lines could be very beneficial when followed. But they needn't be followed. Or those uses of AI could be conjoined with other uses to shortcut thinking. I'm not defeatist, just grateful for my own cultivation of thinking through unassisted writing.
