Purpose of this page | Using this page | ||||||||
---|---|---|---|---|---|---|---|---|---|
This page is for discussion related to student assignments and the Wikipedia Education Program. Please feel free to post, whether you're from a class, a potential class, or if you're a Wikipedia editor. Topics for this board might include:
There are other pages more appropriate for dealing with certain specific issues:
|
Managing threads If you'd like to make sure a thread does not get archived automatically after 30 days, use {{Do not archive until}} at the top of the section. Use {{User:ClueBot III/ArchiveNow}} within a section to have it archived (more or less) immediately. A brief Archives page lists them with the years in which those now inactive discussions took place.
| ||||||||
Index 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
This page has archives. Sections older than 30 days may be automatically archived by ClueBot III when more than 4 sections are present. |
Template:End of course week leaves unclosed div tag
[edit]Template:End of course week leaves an unclosed <div>
, which is causing a plethora of missing end tag lint errors. Please discuss at Template talk:End of course week#Template leaves trailing <div>, no need to discuss it here also. —Anomalocaris (talk) 19:13, 30 May 2025 (UTC)
Best practice for trasnlation assignments?
[edit]Do we have any guides for students on translating articles? I am familiar with general guides (Wikipedia:Translation or User:TheLonelyPather/Essays/Guide for a translator, as well as the excellent, IMHO, meta:OKA/Instructions for editors), but I don't recall one aimed at students.
I am also curious what do you think about the requirement to check translated content (stable on another wiki) for Wikipedia:Close paraphrasing. Recently several of the articles translated by my students (from zh to en) have been found to contain close paraphrasing. To be clear, the students did not add it - they simply translated content from zh wiki, where it was stable for many years, but where also apparently nobody noticed this problem. Do you think the students and/or the instructors should be required to check for close paraphrasing in the articles they select for translation? Or should it be mentioned as a best practice? It is not, as far as I can tell, even suggested as a best practice in the guides linked above. Should it be added to them (or to a guide for students on translations, if we have one and it isn't there already)? Piotrus at Hanyang| reply here 08:03, 21 June 2025 (UTC)
- First, I always like to put in a shameless plug for WP:ASSIGN. For the specific question of close paraphrasing, as I see it, translation does not provide a "pass" for our copyright policy, or for the important teaching principle of not plagiarizing. If the learning experience is going to focus on translation, as opposed to researching and writing from scratch, it seems to me that instructors should insist that students check the zh source material for close paraphrasing, and remove or correct any such content, before beginning to translate. I can also observe, from back in my own days as a university professor (in the US, but teaching many international students), that different parts of the world have very different cultural assumptions about paraphrasing. I had many international students who were surprised that it was a "thing", because they had previously been taught that imitating the instructor or the text book was the best way to demonstrate mastery of the assignment, so they were taken aback when I told the class that copying and close paraphrasing were unacceptable. So my personal opinion is that it's a good educational objective to have students be attentive to it (and thus, learn the concept of credit-where-credit-is-due for scholarly work), as well as being a community expectation on the English Wikipedia that close paraphrasing should not find its way into mainspace. --Tryptofish (talk) 22:46, 21 June 2025 (UTC)
- Piotrus, you have already found the links I would have recommended. As far as close paraphrasing in translation (or copying) as long as the source is a Wikimedia project, they are welcome to copy or translate or copy as exactly or as closely as they wish, as long as it is properly attributed per WP:CWW (per our terms of use §7c) (and as long as the zh-wiki or whatever source is not itself a copy violation). Needless to say, they can't do that from a textbook or any external copyrighted source. Mathglot (talk) 03:22, 7 July 2025 (UTC)
- @Mathglot The question is - whose responsibility is it to determine whether "the zh-wiki or whatever source is not itself a copy violation"? Such determination takes considerable time and requires relatively advanced skills, and AFAIK it is not even recommended as best practice or mentioned at all as something to pay attention to in the translation guides I know. I keep wondering if it should be required or recommended? Some folks at ANI appear to think so (having criticized my class projects for translating stuff from zh wiki that apparently was closely paraphrased there, but not tagged as such, or removed, until some RC patrollers on en wiki checked for this). Piotr Konieczny aka Prokonsul Piotrus| reply here 12:19, 7 July 2025 (UTC)
- I doubt that any discussion here will determine whether or not to change policy in this regard. But there is a clear message in the reactions you apparently had at ANI: the community here won't like it if copyright-violating material finds its way into mainspace here, whatever the path it took to get here, and the blame will be placed on whoever made the edits here, regardless of what might have happened at zh-wiki. --Tryptofish (talk) 23:35, 7 July 2025 (UTC)
- @Mathglot The question is - whose responsibility is it to determine whether "the zh-wiki or whatever source is not itself a copy violation"? Such determination takes considerable time and requires relatively advanced skills, and AFAIK it is not even recommended as best practice or mentioned at all as something to pay attention to in the translation guides I know. I keep wondering if it should be required or recommended? Some folks at ANI appear to think so (having criticized my class projects for translating stuff from zh wiki that apparently was closely paraphrased there, but not tagged as such, or removed, until some RC patrollers on en wiki checked for this). Piotr Konieczny aka Prokonsul Piotrus| reply here 12:19, 7 July 2025 (UTC)
Students and AI slop
[edit]I was asked to bring this here by @Mathglot:, so here goes a straight copy-paste:
Hi Brianda,
I've noticed several students using chatbots to write Wikipedia articles. Sometimes they use chatbot to write content, then later add legit references, but sometimes they use chatbots to add references, references which don't actually exists.
Please coordinate with WikiEd and other teachers/instructors that AI bots are a blight on humanity and should not be used to fulfill assignements, as they are especially damaging to endeavours like Wikipedia.
As for links, the issue is very general and I don't want to single out an individual student here.
Ressources that can be used to check if a student used AI to write something
- https://gptzero.me/ [update: or other AI-detection tools] Insert student text and it'll come up with a likelyhood that it's AI-written. Probably the best method I know.
- WP:UPSD, checks for AI-selected URLs amongst other things (see WP:UPSD#AI-generated). This requires low effort, but will miss lots of things.
- In general, short "bullet list" are very often indicative of AI slop.
Thanks. Headbomb {t · c · p · b} 05:24, 7 July 2025 (UTC)
- I have been hearing anecdotally from (non-Wiki Ed) teachers/professors who have been reevaluating their approach to giving homework assignments to their students, or even abandoning them. I wonder how this is affecting instructors enrolled with Wikipedia Education, and the program as a whole and any plans about Wiki Ed adjusting to the new reality. Mathglot (talk) 06:26, 7 July 2025 (UTC)
- Students tend to believe we can't tell, so I doubt that instructors telling their students not to use them will have much effect, but I agree that WikiEd should warn instructors to tell students that they're very likely to get blocked from editing if they use AI. -- asilvering (talk) 06:32, 7 July 2025 (UTC)
- Or just have the content deleted. Graeme Bartlett (talk) 07:01, 7 July 2025 (UTC)
- I agree that TNT is the best approach -- Guerillero Parlez Moi 08:04, 7 July 2025 (UTC)
- At my university, we received instructions that instructors and TAs cannot use AI detectors as evidence to determine whether the text is likely to be written by AI. The rationale we got is that they are not very reliable and can lead to false accusation. According to a recent article in ZDNET, GPTZero is accurate 80% of the time but the accuracy is declining over time. I do not recommend including a AI-checker as a resource. OhanaUnitedTalk page 13:32, 7 July 2025 (UTC)
- I don't think a conclusion like "GPTZero is accurate 80% of the time" can be extrapolated from that data. That was a single failure, in a test of five texts. I do agree that caution is warranted, though. I tend to use checkers as a backup for when I'm pretty sure something is AI (ie, a lower bar than "virtually certain"). If they return any doubt, I give benefit of the doubt. -- asilvering (talk) 13:55, 7 July 2025 (UTC)
- I updated the above to make it clear I'm using gtpzero as an exemple, not as the definite arbiter of thruth, freedom, and liberty. Headbomb {t · c · p · b} 14:44, 7 July 2025 (UTC)
- I simply do not think any AI checkers should be recommended at this point because their reliability are undetermined at this point. OhanaUnitedTalk page 15:59, 7 July 2025 (UTC)
- I updated the above to make it clear I'm using gtpzero as an exemple, not as the definite arbiter of thruth, freedom, and liberty. Headbomb {t · c · p · b} 14:44, 7 July 2025 (UTC)
- I don't think a conclusion like "GPTZero is accurate 80% of the time" can be extrapolated from that data. That was a single failure, in a test of five texts. I do agree that caution is warranted, though. I tend to use checkers as a backup for when I'm pretty sure something is AI (ie, a lower bar than "virtually certain"). If they return any doubt, I give benefit of the doubt. -- asilvering (talk) 13:55, 7 July 2025 (UTC)
- At my university, we received instructions that instructors and TAs cannot use AI detectors as evidence to determine whether the text is likely to be written by AI. The rationale we got is that they are not very reliable and can lead to false accusation. According to a recent article in ZDNET, GPTZero is accurate 80% of the time but the accuracy is declining over time. I do not recommend including a AI-checker as a resource. OhanaUnitedTalk page 13:32, 7 July 2025 (UTC)
- I agree that TNT is the best approach -- Guerillero Parlez Moi 08:04, 7 July 2025 (UTC)
- Or just have the content deleted. Graeme Bartlett (talk) 07:01, 7 July 2025 (UTC)
Instructor who would benefit from WikiEd
[edit]There's a discussion at Wikipedia:Help desk#Adding and editing content on fishes of the world, started by User:N. Mercedis, a faculty member at UC Davis. The instructor and their students would clearly benefit from working with WikiEd, so I hope that someone from there will reach out to the instructor. Thanks! --Tryptofish (talk) 23:40, 7 July 2025 (UTC)