My college kid just got an F because AI checker said his report was AI written

[lewelma]

His group got a 50% on their report with no other comment than the AI checker said it was likely written by AI.  The 3 kids doing the project are all top kids, with A averages in a school that has no grade inflation, so the average grade is a C+/B-. None of them used AI. This was for a GIS class and they did use the online instruction manual for ArcGIS Pro to get a lot of their insightful ideas.  The professor did grade them, not a TA. And also gave them 100% for their presentation that was based on the report. They did NOT get an email about academic dishonesty from the University. So what should he do?

 

[prairiewindmomma]

Is there a log of the version history? Some programs (like google docs and microsoft word) have this. This false flag issue is a not uncommon problem from what I am hearing from my college kid.

I’d make an appointment with the prof and bring in the log version, the research data, online search histories, handwritten notes, and whatever other documentation the student has to discuss the matter.

[lewelma]

His friend just ran it through 3 different AI checkers.  One is what the prof found likely AI written, one said it is not AI written at all, and one said the title was AI written. So I can't figure out what the prof can stand on. My kid is in the middle of exams, so not great timing to be dealing with this nonsense.

[lewelma]

6 minutes ago, prairiewindmomma said:

 

I’d make an appointment with the prof and bring in the log version, the research data, online search histories, handwritten notes, and whatever other documentation the student has to discuss the matter.

Thanks for that. I'll see what they have. My son has dysgraphia, so wouldn't have any handwritten notes. No printer either. So I'll suggest the log and the online search history.

 

Edited June 12 by lewelma

[lewelma]

Ah, he can't get his search history because his computer died last week and he is using  a friend's computer for all his exams. 

[Bootsie]

He should begin by talking to the professor to get an idea of exactly what the professor's concerns regarding the paper are.  Whether I would suggest he do this by himself or with his team members would depnd somewhat on how they went about writing the paper and how well he knwos the other students.  If they all worked together on the entire paper, it would be a good idea to go in together.  If the tag-teamed the work, it might be best to go alone.  

Some schools allow for a professor to lower a grade for academic dishonesty issues and it is up to the student to appeal the professor's decision.  Other schools require a professor to present concerns about academic dishonesty to an administrator who then follows a process to determine whether academic dishonesty has occurred or not.  I would suggest that he become familiar with the process at his university in case the situation is not resolved after meeting with the professor. 

[Mom_to3]

Do the other students on the project have either search histories, or earlier drafts of the project (or e-mails/texts) that they've exchanged? My (much younger) kids save multiple versions of their drafts for writing assignments/projects - and I imagine would send these off to the teacher immediately if they ever raised doubts about AI.

[lewelma]

Thanks for the ideas. He and his friends will be dealing with this asap. Interestingly, my university does not allow professors to run stuff through AI detection software as it is so inaccurate. So I'm not sure of the policy at my ds's school. I'll have him look it up.

[Laura Corin]

4 hours ago, lewelma said:

His group got a 50% on their report with no other comment than the AI checker said it was likely written by AI.  The 3 kids doing the project are all top kids, with A averages in a school that has no grade inflation, so the average grade is a C+/B-. None of them used AI. This was for a GIS class and they did use the online instruction manual for ArcGIS Pro to get a lot of their insightful ideas.  The professor did grade them, not a TA. And also gave them 100% for their presentation that was based on the report. They did NOT get an email about academic dishonesty from the University. So what should he do?

 

In the university where I work, that case would go to a board made up of academics.  The students concerned would be invited to provide evidence (for example notes that formed the background to the piece of work, first drafts, etc.).  I would look into appealing.

[gardenmom5]

Do protest it.

in high school, 1ds was accused to plagiarism by his English teacher because "teenagers don't talk like that". . . . 
basically, 1ds had a better vocabulary than the teacher.    and knew how to use semicolons . .   

Edited June 12 by gardenmom5

[LuvToRead]

2 hours ago, gardenmom5 said:

Do protest it.

in high school, 1ds was accused to plagiarism by his English teacher because "teenagers don't talk like that". . . . 
basically, 1ds had a better vocabulary than the teacher.    and knew how to use semicolons . .   

My oldest had a similar situation in college.  She was not actually accused of using AI, but a professor did tell her certain words she used in a paper were red flags for an AI written paper. She then made a point of using "big words" in front of her professors so they would know she actually does have an extensive vocabulary (at least more so than her peers).  

It would be worth talking to the professor.

[pinball]

7 hours ago, lewelma said:

His group got a 50% on their report with no other comment than the AI checker said it was likely written by AI.  The 3 kids doing the project are all top kids, with A averages in a school that has no grade inflation, so the average grade is a C+/B-. None of them used AI. This was for a GIS class and they did use the online instruction manual for ArcGIS Pro to get a lot of their insightful ideas.  The professor did grade them, not a TA. And also gave them 100% for their presentation that was based on the report. They did NOT get an email about academic dishonesty from the University. So what should he do?

 

Wow, that’s maddening. 
 

I think I’d tell my kid to immediately write an email to the professor and CC other people like my advisor, head of the department that professor works under and maybe the VP of Academics. 
 

the purpose would be to get on record that he is denying the accusation and give a basic outline of the work he did.

[Clemsondana]

These situations are hard for both students and faculty.  I don't use AI checkers because I don't think they are reliable, but I've had a couple of students that I suspect were cheating their way through my class.  My usual indicator is when a student suddenly starts giving advanced answers that include material not covered in class while simultaneously being unable to answer easy questions that are specific to the class (in lab last week, what was?...) or leaving every chart, table, or crossword puzzle blank (AI can't do those).  Whenever I suspect cheating, I generally give 0 points on that question and add a note that this answer doesn't seem to match the other types of answers, but I'd be happy to talk with the student about it.  If I have multiple such answers on the same assignment then I assign a zero to the whole assignment and say that I'm happy to talk with the student to sort it out.  It's possible that's what the professor is doing - he suspects use of AI but if the students can have an intelligent discussion with him about the topic then he'll believe that they wrote it.  In my case, I've never had a student in this situation who was willing to have a conversation, and most admit that they used some type of online resource.  I would be thrilled for a student to have a chat about their work - I'd much rather assign As than Fs.  I'd recommend not going in confrontationally.  If there is support, such as updates to the file, then take that, but be prepared for it to go well.  I absolutely hate dealing with cheating, and most semesters I spend more time on it than I do on grading.  It's possible that the prof will be difficult, but it's also possible that he's just set 'do with the checker says' as the default but is very willing to have it proven wrong.  

[HomeAgain]

One thing he should ask the professor is how it deviates from the quality and consistency of past work.  The students should be prepared with work samples to show, if necessary, that the work produced is not out of the norm, nor was the presentation on the topic inadequately prepared for.

[maize]

These AI checkers are completely unreliable.

I'd look for something written by the professor and run that through AI checkers, then show the results to the professor.

Edited June 12 by maize

[maize]

Alternatively run some of your own papers from pre-ChatGPT years through AI checkers.

[gardenmom5]

1 hour ago, maize said:

These AI checkers are completely unreliable.

I'd look for something written by a professor and run that through AI checkers, then show the results to the professor.

preferably - something from THAT professor.

[Miss Tick]

I don't think any professor thinks AI checkers are wholly reliable. Start by suggesting your ds approach the professor to find out what the appeal process is.

DH has received papers from senior level students that he didn't have to screen with AI checkers because the student had not removed the "Your question was ..., ChatGPT responded ...." header. So using an AI checker is, sadly, a completely reasonable first step in grading a paper. Why spend your time checking a paper the student didn't spend time writing?

We know OP's ds wrote this paper, partly because we have watched him grow up for years. This professor, however, may have just met him this year and doesn't know him the way we do.

[Bootsie]

3 hours ago, pinball said:

I think I’d tell my kid to immediately write an email to the professor and CC other people like my advisor, head of the department that professor works under and maybe the VP of Academics. 
 

the purpose would be to get on record that he is denying the accusation and give a basic outline of the work he did.

I would advise against emailing other people as a starting point.  First, the student should meet with the professor to see exactly what the professor's concerns are.   There may be a simple resolution.  The student will also know exactly what needs to be addressed/countered.  A specific process for appeal should be outlined in the university handbook.  That process needs to be followed, if after talking to the professor, an appeal is needed.  Emailing information immediately to people outside of that protocol could interfer with the process.  

[Heartstrings]

If they used Grammarly that can kick back as being AI written.  I think it’s ridiculous but it’s definitely a thing.   It was all over TikTok not long ago.  What can be done varies by university.  The one that was viral a while ago the student just got screwed basically, she was put in academic probation despite a letter written from Grammarly itself and proof that the school website itself promotes the use of Grammarly.   
 

This is going to be a really big problem going forward.

[pinball]

1 hour ago, Bootsie said:

I would advise against emailing other people as a starting point.  First, the student should meet with the professor to see exactly what the professor's concerns are.   There may be a simple resolution.  The student will also know exactly what needs to be addressed/countered.  A specific process for appeal should be outlined in the university handbook.  That process needs to be followed, if after talking to the professor, an appeal is needed.  Emailing information immediately to people outside of that protocol could interfer with the process.  

I appreciate your advice but in my life, I’d still start with an email and CC the people I said.

Bc the same could be said of the professor…there could have been a simple resolution if he thought the paper was written by AI. He could have told the students his concerns before failing them outright.

I think it’s a nasty way to treat people, especially if the students had other good solid work in the class. 

[Heartstrings]

47 minutes ago, pinball said:

ink it’s a nasty way to treat people, especially if the students had other good solid work in the class.

It’s nasty and I can’t get over how LAZY it all is.   “Ooh, my crappy free website says you used AI, I can’t possibly argue with that.”    

[wintermom]

All the students who were accused of using AI should definitely go speak to the dean of their department - all together. The students do have power. They need to bring this to the attention to the right people - and that isn't necessarily the prof. The prof has already made his decision, and what use is it going to be to complain to him? It's the deptartment heads that take in the student feedback and then go deal with the prof. If the complaints are legit, the prof will get told they need to change things.

The student union may also have some support for these kinds of issues. That is another body that will help support student rights on campus. This AI issue is likely to cause problems not just for this one class. Back during COVID lockdowns, when university exams were done online with proctoring software, there were lots of issues with some of the cameras not being able to recognize very dark skinned students. Student unions at some universities were instrumental in helping to support the students by amplifying the voices of many individuals to make the broader issue recognized.

Edited June 12 by wintermom

[lewelma]

Thanks guys for all the ideas.  I slept really badly last night and I'm sure my DS did too. This is exam week and he simply doesn't have time to deal with this. His exams are worth 50% of his grade.

I think there are so many problems with this, but 2 stand out. 

1) the entire group got and F even though only 1 section was flagged as AI written. We know it wasn't (kid is a top student that DS knows well). But moving forward on group projects, this is a problem if DS gets put in a group with someone who uses AI.

2) I do wonder if this isn't going to happen over and over again to DS, which is crazy unfair.

 

[lewelma]

2 hours ago, Heartstrings said:

If they used Grammarly that can kick back as being AI written

 

Prof in his comment said "it could be Grammarly or ChatGBT but Im guessing the latter."  That was all the reason that was given for 50%.

[lewelma]

6 hours ago, Clemsondana said:

These situations are hard for both students and faculty.  I don't use AI checkers because I don't think they are reliable, but I've had a couple of students that I suspect were cheating their way through my class.  My usual indicator is when a student suddenly starts giving advanced answers that include material not covered in class while simultaneously being unable to answer easy questions that are specific to the class (in lab last week, what was?...) or leaving every chart, table, or crossword puzzle blank (AI can't do those).  Whenever I suspect cheating, I generally give 0 points on that question and add a note that this answer doesn't seem to match the other types of answers, but I'd be happy to talk with the student about it.  If I have multiple such answers on the same assignment then I assign a zero to the whole assignment and say that I'm happy to talk with the student to sort it out.  It's possible that's what the professor is doing - he suspects use of AI but if the students can have an intelligent discussion with him about the topic then he'll believe that they wrote it.  In my case, I've never had a student in this situation who was willing to have a conversation, and most admit that they used some type of online resource.  I would be thrilled for a student to have a chat about their work - I'd much rather assign As than Fs.  I'd recommend not going in confrontationally.  If there is support, such as updates to the file, then take that, but be prepared for it to go well.  I absolutely hate dealing with cheating, and most semesters I spend more time on it than I do on grading.  It's possible that the prof will be difficult, but it's also possible that he's just set 'do with the checker says' as the default but is very willing to have it proven wrong.  

 

This paper was written over a month ago on highly technical detail. DS and group are meeting with the prof at noon. I hate to have him go study his own paper so he can talk to it but sounds like he should. Main problem is that the paper is only worth 10% of his grade, and the exams he has in 2 days is worth 50%.  I think DS feels it is the principle of the matter, but it could screw his grade in a different class. 

[lewelma]

3 hours ago, Miss Tick said:

 

 

We know OP's ds wrote this paper, partly because we have watched him grow up for years. This professor, however, may have just met him this year and doesn't know him the way we do.

 

Thanks for this. DS's dysgraphia meant that he and I sat on the sofa for 3 hours per day for 6 years remediating his writing. He prides himself, and even identifies with the fact that he can write. Because his writing represents who he is and how he was not willing to give up. This whole thing is a slap in the face. 

[lewelma]

1 hour ago, wintermom said:

All the students who were accused of using AI should definitely go speak to the dean of their department - all together. The students do have power. They need to bring this to the attention to the right people - and that isn't necessarily the prof. The prof has already made his decision, and what use is it going to be to complain to him? It's the deptartment heads that take in the student feedback and then go deal with the prof. If the complaints are legit, the prof will get told they need to change things.

 

Regardless of the outcome of talking to the prof, DS is going to bring it to the head of the department, who he knows.  The department needs to be aware that its top students are being targeted. 

 

 

I think my biggest complaint is that the prof did not even bring them in to talk to them about it before giving them an F. He did give them 100% on the presentation that they did on the paper. So he knows that they made those slides, practiced, and could talk to the content. It makes me pretty mad.

[QueenCat]

31 minutes ago, lewelma said:

Regardless of the outcome of talking to the prof, DS is going to bring it to the head of the department, who he knows.  The department needs to be aware that its top students are being targeted. 

 

If the professor agrees that he/she made an error & corrects the grade, and your son will have this professor again, I would not go to the department head. That will create further conflict with the professor.

[lewelma]

Just now, QueenCat said:

If the professor agrees that he/she made an error & corrects the grade, and your son will have this professor again, I would not go to the department head. That will create further conflict with the professor.

I hear that, but if this is going to happen again and again I do think someone has to ask the department how they will handle it in the future. Perhaps he could not mention the professor's name and just write the email laying out the issues with what has happened .  He is not likely to have the prof again, but NZ is a small village, and everyone knows everyone. Thoughts?

[QueenCat]

1 minute ago, lewelma said:

I hear that, but if this is going to happen again and again I do think someone has to ask the department how they will handle it in the future. Perhaps he could not mention the professor's name and just write the email laying out the issues with what has happened .  He is not likely to have the prof again, but NZ is a small village, and everyone knows everyone. Thoughts?

I would listen to what some of the professors/instructors are saying. Like @Bootsie and @regentrude. There are a couple of others, as well.

[Miss Tick]

16 minutes ago, lewelma said:

I hear that, but if this is going to happen again and again I do think someone has to ask the department how they will handle it in the future. Perhaps he could not mention the professor's name and just write the email laying out the issues with what has happened .  He is not likely to have the prof again, but NZ is a small village, and everyone knows everyone. Thoughts?

You don't have any evidence that it will happen "again and again". Perhaps hold off until after the meeting at least before taking any actions.

I know what you mean about "a slap in the face", though. The self-checkout machine accused me of shoplifting not long ago and I felt the same way. AI can be a ... rude, inconsiderate despot.

[wintermom]

1 hour ago, lewelma said:

Regardless of the outcome of talking to the prof, DS is going to bring it to the head of the department, who he knows.  The department needs to be aware that its top students are being targeted. 

 

 

I think my biggest complaint is that the prof did not even bring them in to talk to them about it before giving them an F. He did give them 100% on the presentation that they did on the paper. So he knows that they made those slides, practiced, and could talk to the content. It makes me pretty mad.

It sounds like your ds has a plan of action. I really hope it works out well.  Hopefully you can keep us posted on the outcome.

I'd be very interested to hear about how these early AI issues are being dealt with in report writing. It's a HUGE issue, and not just for colleges. In the wider research field, there are privacy issues. Some AI generated websites keep your original material and add it to their own data banks. If you're working with an REB, you'd better know that your data is secure.

If this prof inputted his students' writing into an AI site, then he has potentially taken away their choice of privacy. It wasn't his property. There could well be a legal issue here.

 

Edited June 12 by wintermom

[Heartstrings]

1 hour ago, wintermom said:

Some AI generated websites keep your original material and add it to their own data banks. If you're working with an REB, you'd better know that your data is secure.

If this prof inputted his students' writing into an AI site, then he has potentially taken away their choice of privacy. It wasn't his property. There could well be a legal issue here.

This seems especially likely if he is using a free checker.  If you aren't paying for the service online, YOU are the product being sold.    Not to mention that then the AI detector can start detecting plagiarism because it's comparing a student's new work with an older work of that same student that was previously scanned into the detector.   

 

This whole thing is going to be such a mess and I can't believe universities aren't figuring out how to get ahead of this before it all blows up. 

[wintermom]

42 minutes ago, Heartstrings said:

This seems especially likely if he is using a free checker.  If you aren't paying for the service online, YOU are the product being sold.    Not to mention that then the AI detector can start detecting plagiarism because it's comparing a student's new work with an older work of that same student that was previously scanned into the detector.   

 

This whole thing is going to be such a mess and I can't believe universities aren't figuring out how to get ahead of this before it all blows up. 

If the prof included full names, he's a complete idiot. Sound like he was and idiot anyway, but sheesh!

Edited June 12 by wintermom

[Mrs Tiggywinkle Again]

I mean one of my students turned this in so….

We use turnitin at my college but no specific AI checker as they are notoriously unreliable. 

[lewelma]

He met with the prof. It took them a long time to convince him that they had written it without AI. Apparently, their level of writing was too high. But they showed him the outline, the draft paragraphs (with comments to self and question marks), the addition of references over time, etc. He apparently went over multiple drafts quite carefully. But in the end he changed the grade from 50% to 100%! 

 

Apparently last year in the sophomore class (ds is a junior), he had EIGHT groups (so 24 people) who he found had used AI, and none of them contested it. Six groups rewrote it (which he allowed and which is why he did not turn it in for academic dishonesty) and two groups accepted the 50%. He said he has never had a group contest the accusation of using AI in the 2 semesters he has used the AI checker for any of his classes. By the end he was very apologetic and said that in the future he would be more careful. So all good in the end.

DS will be talking to the head of the department next term (we are only half way through the year here in the southern hemisphere). This will not be about the professor, but about the policy or lack there of and the ramification to good writers.  Trying to decide if this should be done by email or face to face.

 

 

Edited June 13 by lewelma

[bookbard]

14 minutes ago, lewelma said:

He met with the prof. It took them a long time to convince him that they had written it without AI. Apparently, their level of writing was too high. But they showed him the outline, the draft paragraphs (with comments to self and question marks), the addition of references over time, etc. He apparently went over multiple drafts quite carefully. But in the end he changed the grade from 50% to 100%! 

 

Apparently last year in the sophomore class (ds is a junior), he had EIGHT groups (so 24 people) who he found had used AI, and none of them contested it. Six groups rewrote it (which he allowed and which is why he did not turn it in for academic dishonesty) and two groups accepted the 50%. He said he has never had a group contest the accusation of using AI in the 2 semesters he has used the AI checker for any of his classes. By the end he was very apologetic and said that in the future he would be more careful. So all good in the end.

DS will be talking to the head of the department next term (we are only half way through the year here in the southern hemisphere). This will not be about the professor, but about the policy or lack there of and the ramification to good writers.  Trying to decide if this should be done by email or face to face.

 

 

I know how hard you've worked with your son over the years. I want to say congratulations to both of you. This is a very good outcome, not just for him, but hopefully for the entire university. I think it is very wrong (in fact it infuriates me) that people are putting other people's work into AI programs. I also know a lot of academics and they are just not thinking super hard about all of this. It's because they're overworked and time poor, I know, but this is very serious, and it is going to get even more serious, very soon. Once again, congratulations to you and your son. 

[lewelma]

23 minutes ago, bookbard said:

I know how hard you've worked with your son over the years. I want to say congratulations to both of you. This is a very good outcome, not just for him, but hopefully for the entire university. I think it is very wrong (in fact it infuriates me) that people are putting other people's work into AI programs. I also know a lot of academics and they are just not thinking super hard about all of this. It's because they're overworked and time poor, I know, but this is very serious, and it is going to get even more serious, very soon. Once again, congratulations to you and your son. 

My department does not use AI checkers at all. They have decided it is just too fraught. 

At ds's university, It will be interesting to see what the head of the department says. My DS has gotten 2 As and and A+ in this professor's classes (this is for a university in NZ without grade inflation). So to have a student like this accused will be eye opening and hopefully lead to some change that will help others in the future. 

Edited June 13 by lewelma

[lewelma]

23 minutes ago, bookbard said:

I know how hard you've worked with your son over the years. I want to say congratulations to both of you.

 

Thanks for this. We did work very very hard.

Being 2E, he has now won multiple awards for his writing while concurrently needed a scribe to fill in scantron bubbles for him. He still is basically unable to physically write as he cannot remember how to form the letters. 

[Eos]

4 hours ago, lewelma said:

face to face.

Seems better to me since it's about navigating human vs machine. Just a thought - I know you and he will choose the right way, considering his 2e needs. Good on him for following up, what a shock this must have been and a scary thing to be accused of.

[LuvToRead]

I am so happy for the good outcome! You must be so proud of your son!  

[Clemsondana]

I'm glad it went well, and also glad that the professor explained how common this is.  People are quick to blame instructors when they automatically assign poor grades, but it's not an exaggeration to say that more than 85% of the time (in this professor's case, and it's 100% for many instructors) the zero is the correct grade to assign.  I, too, have never had a student contest an AI query.  I've found that, without assigning a zero, the students don't reply to my questions and ignore the comments.  My spouse taught a college coding class pre-AI had and had a semester where at least 1/2 of the class cheated in some way.  In that case, he was able to prove it because he could look at their files as they uploaded in the computer lab and see that they went from having no code to having complete code between updates, and their comments on the code were identical.  Many of us reach the point where our entire class structure is being determined by trying to ensure that students actually do enough work to learn something.  Academics are discussing moving to a model where everything is done in class - old-school blue book style.  But that's a terrible model for students who need to learn to write long papers or do big coding projects.  Students have always cheated, but with AI it got the deadly combination of being easy to do and very hard to detect.  There is a lot of agonizing about how to fight this without causing undue stress and angst for good students.  I'm not sure that it's possible.  I hate that this was done so abruptly, but I'm also pretty sure that even if the prof had included (like I do) a note that he'd be happy to chat about it, the students wouldn't have been any less stressed.  

I would be careful in how this is approached with the dean, if he feels that he has to go that route - once something is resolved to the satisfaction of both the instructor and student, causing that faculty member to have to deal with meetings with the dean has the potential to cause conflict.  Or, put another way, if you worked to resolve a problem, even if it was due to your own mistake (and I'm not saying this was), and somebody complained to your boss after it was fixed, how happy would you be to go the extra mile with that client in the future?  If a formal process is included, it will likely harm students because it take the professor's flexibility. Currently students get one 'free pass' where they can rewrite.  If a professor 'gets in trouble' for the current process, a potential outcome is that faculty will start turning in all instances of suspected cheating to an honor panel rather than risk making a judgement call about how to handle it again.  If that happens, even if it is favorably resolved, it will usually be a much more time consuming process for everybody that drags on for weeks and likely won't serve the students as well as a quick meeting with the prof did. 

Academic departments and colleges are struggling with how to deal with this.  As I said, it's also not an exaggeration to say that I spend more time trying to track down cheating problems than I do grading.  With AI it's absolutely exasperating because there is no way to check for sure.  And, it's everywhere.  Grammerly has added an AI component.  Last year, we finally confronted one student because the cheating was rampant across multiple online classes.  Their first response was 'I just use Grammarly' but it turns out that they were using the part of Grammarly where you can type a question in and it can generate an answer.  They had explained it as 'Grammarly just adds stuff sometimes' but when one of the teachers explored, it turns out that it will indeed write the essay for you.  My most recent indicator of AI is in article reviews.  I assign students to write 3 short (1-2 paragraph) summaries of articles about science.  Most students grab something from the news, and others pick articles about areas of interest (pitching injuries, horses, workout supplements).  I've had a handful who suddenly started submitting writeups about primary journal articles.  It seemed crazy - most high schoolers aren't going to read a 10 page journal article about a biochemical pathway when their classmates are reading the 1-page news article about the new species of lizard found in Madagascar.  It can happen, but is unlikely, especially for a struggling student.  It turns out that AI doesn't care what article it summarizes.  It's another instance where 'I'm happy to talk about this with you' would be really informative.  I'd love to chat with a student who actually read a primary article, but it has never happened.  

Edited June 13 by Clemsondana

[wintermom]

1 hour ago, Clemsondana said:

I'm glad it went well, and also glad that the professor explained how common this is.  People are quick to blame instructors when they automatically assign poor grades, but it's not an exaggeration to say that more than 85% of the time (in this professor's case, and it's 100% for many instructors) the zero is the correct grade to assign.  I, too, have never had a student contest an AI query.  I've found that, without assigning a zero, the students don't reply to my questions and ignore the comments.  My spouse taught a college coding class pre-AI had and had a semester where at least 1/2 of the class cheated in some way.  In that case, he was able to prove it because he could look at their files as they uploaded in the computer lab and see that they went from having no code to having complete code between updates, and their comments on the code were identical.  Many of us reach the point where our entire class structure is being determined by trying to ensure that students actually do enough work to learn something.  Academics are discussing moving to a model where everything is done in class - old-school blue book style.  But that's a terrible model for students who need to learn to write long papers or do big coding projects.  Students have always cheated, but with AI it got the deadly combination of being easy to do and very hard to detect.  There is a lot of agonizing about how to fight this without causing undue stress and angst for good students.  I'm not sure that it's possible.  I hate that this was done so abruptly, but I'm also pretty sure that even if the prof had included (like I do) a note that he'd be happy to chat about it, the students wouldn't have been any less stressed.  

I would be careful in how this is approached with the dean, if he feels that he has to go that route - once something is resolved to the satisfaction of both the instructor and student, causing that faculty member to have to deal with meetings with the dean has the potential to cause conflict.  Or, put another way, if you worked to resolve a problem, even if it was due to your own mistake (and I'm not saying this was), and somebody complained to your boss after it was fixed, how happy would you be to go the extra mile with that client in the future?  If a formal process is included, it will likely harm students because it take the professor's flexibilitp. Currently students get one 'free pass' where they can rewrite.  If a professor 'gets in trouble' for the current process, a potential outcome is that faculty will start turning in all instances of suspected cheating to an honor panel rather than risk making a judgement call about how to handle it again.  If that happens, even if it is favorably resolved, it will usually be a much more time consuming process for everybody that drags on for weeks and likely won't serve the students as well as a quick meeting with the prof did. 

Academic departments and colleges are struggling with how to deal with this.  As I said, it's also not an exaggeration to say that I spend more time trying to track down cheating problems than I do grading.  With AI it's absolutely exasperating because there is no way to check for sure.  And, it's everywhere.  Grammerly has added an AI component.  Last year, we finally confronted one student because the cheating was rampant across multiple online classes.  Their first response was 'I just use Grammarly' but it turns out that they were using the part of Grammarly where you can type a question in and it can generate an answer.  They had explained it as 'Grammarly just adds stuff sometimes' but when one of the teachers explored, it turns out that it will indeed write the essay for you.  My most recent indicator of AI is in article reviews.  I assign students to write 3 short (1-2 paragraph) summaries of articles about science.  Most students grab something from the news, and others pick articles about areas of interest (pitching injuries, horses, workout supplements).  I've had a handful who suddenly started submitting writeups about primary journal articles.  It seemed crazy - most high schoolers aren't going to read a 10 page journal article about a biochemical pathway when their classmates are reading the 1-page news article about the new species of lizard found in Madagascar.  It can happen, but is unlikely, especially for a struggling student.  It turns out that AI doesn't care what article it summarizes.  It's another instance where 'I'm happy to talk about this with you' would be really informative.  I'd love to chat with a student who actually read a primary article, but it has never happened.  

I believe that there needs to be a lot more consultation between the university/college leadership, professors and the students (both TAs and those taking courses) regarding the way forward. If a university/college does not have a standard policy on AI checker that they are informing students about, then they really should. There are implications around privacy and intellectual ownership, and students should have a say in what is being done with their property.

My dd just informed me that her university has a policy where professors use an AI checker for assignments, and that this information is given to the students in the course syllabus. Dd didn't know what type of AI checker it was, and what happened to the content inputted. At least the student were informed AHEAD of time. 

Edited June 13 by wintermom

[maize]

Kudos to your son and his group-mates for making their case, and kudos to the professor for considering it thoughtfully.

This is a hard new challenge for everyone to navigate.

[Clemsondana]

2 minutes ago, wintermom said:

I believe that there needs to be a lot more consultation between the university/college leadership, professors and the students (both TAs and those taking courses) regarding the way forward. If a university/college does not have a standard policy on AI checker that they are informing students about, then they really should. There are implications around privacy and intellectual ownership, and students should have a say in what is being done with their property.

My dd just informed me that her university has a policy where professors use an AI checker for assignments, and that this information is given to the students in the course syllabus. Dd didn't know what type of AI checker it was, and what happened to the content inputted. At least the student were informed AHEAD of time. 

I don't have a feel for how much any given university has worked with students on their policies.  I know that some faculty have boilerplate in their syllabi about using Turnitin, and I'd assume that some have similar info about using AI checkers and/or that they tell students they are using it.  I don't think that students have any better feel for how to handle this than the instructors do.  Some students, especially those that put a lot of effort into their work, want schools to come down like a ton of bricks on anybody caught using AI because they don't think it's fair to be competing against it.  Others don't see why it's a problem - they assume that AI will be available to them on the job, so why shouldn't they use it?  When I was teaching students to calculate dose years ago, I had students argue that they shouldn't have to learn that - as nurses, they'd use an app.  It's the same thing, only applying to everything.  

One of the saving graces is that, right now, AI isn't always trained correctly.  Depending on how you word the question, it can give really bad answers, or include made-up information.  I'd hate to think that the work run through a checker is being used to train AI, since student work is often a bit questionable.  I'm not sure what the privacy concerns about using an AI checker are that are different from using Turnitin, which has been used for at least a decade, but I haven't looked into it closely.  

[wintermom]

5 minutes ago, Clemsondana said:

I don't have a feel for how much any given university has worked with students on their policies.  I know that some faculty have boilerplate in their syllabi about using Turnitin, and I'd assume that some have similar info about using AI checkers and/or that they tell students they are using it.  I don't think that students have any better feel for how to handle this than the instructors do.  Some students, especially those that put a lot of effort into their work, want schools to come down like a ton of bricks on anybody caught using AI because they don't think it's fair to be competing against it.  Others don't see why it's a problem - they assume that AI will be available to them on the job, so why shouldn't they use it?  When I was teaching students to calculate dose years ago, I had students argue that they shouldn't have to learn that - as nurses, they'd use an app.  It's the same thing, only applying to everything.  

One of the saving graces is that, right now, AI isn't always trained correctly.  Depending on how you word the question, it can give really bad answers, or include made-up information.  I'd hate to think that the work run through a checker is being used to train AI, since student work is often a bit questionable.  I'm not sure what the privacy concerns about using an AI checker are that are different from using Turnitin, which has been used for at least a decade, but I haven't looked into it closely.  

This is where a lot more deep investigation needs to be done, I think, but that's my job - evaluating processes. 😉  Using apps is only as secure as the quality of the apps; using AI is only as secure as the quality of the AI. Who is checking that? Why would a nurse rely on an app when the consequences of a error could be life or death? That is damn scary. I hope it's not a common thing. Covering my ears and running....

[Clemsondana]

12 minutes ago, wintermom said:

This is where a lot more deep investigation needs to be done, I think, but that's my job - evaluating processes. 😉  Using apps is only as secure as the quality of the apps; using AI is only as secure as the quality of the AI. Who is checking that? Why would a nurse rely on an app when the consequences of a error could be life or death? That is damn scary. I hope it's not a common thing. Covering my ears and running....

It was the day that a small group came to me and said 'We've been talking and we agree that we shouldn't have to learn this...'.  It was 15 years ago...I was teaching at a CC and we used a common syllabus for classes and labs, so even if I had agreed (which I didn't) I had limited ability to make major changes to the class content.  A few weeks later I told them about an 'Untold stories of the ER' episode that I had seen where they went to give a hemorrhaging postpartum woman a shot of Mg and it was missing from the cart, so somebody did a back-of-the-envelope calculation and figured that if they squeezed 1/2 of a drip bag of dilute Mg solution in they would get the equivalent amount into the patient and save her...I told them there was no app for that!  My worry with student use of AI is that they will ultimately know nothing.  Innovation comes from people putting ideas together in novel ways.  If nobody has the base knowledge, then all that they can do is what they are told, and the innovation will be coming from somebody who actually understands what is going on.  There is no app for thinking...it can only respond to inputs.

Edited June 13 by Clemsondana

[PaxEtLux]

I'm very happy to hear that this got resolved.  I suspect the skill of politely working with the professor to correct wrongs is a much more valuable skill to have in the long run than whatever was taught in the class.

And this is going to be a difficult problem for schools in the future.  Clearly it is wrong to hand in, verbatim, a complete paper written by AI. But, where is the line?  Today, everyone would say that you should be able to use google to search for sources for a paper.  Seems like using AI to find sources and help put together your own ideas could be OK, and similar to using google.

But I really must object to the idea that anything could with any degree of certainty could detect that the mere title was AI generated.  The title is usually so short, it would seem that false positives would happen all the time.

[TechWife]

On 6/12/2024 at 12:38 PM, pinball said:

I appreciate your advice but in my life, I’d still start with an email and CC the people I said.

Bc the same could be said of the professor…there could have been a simple resolution if he thought the paper was written by AI. He could have told the students his concerns before failing them outright.

I think it’s a nasty way to treat people, especially if the students had other good solid work in the class. 

I don’t think that it’s reasonable to expect a professor to reach out to schedule a 1:1 at the end of the semester when he’s reading papers and writing & grading finals for multiple classes. I doubt there’s time to meet with every student that wrote a  paper that has issues. It’s the professors job to assess the performance, it’s not “nasty.”

Using the appropriate appeal process is the way to go - not to try to jump over someone’s head. That’s how the world works. You start with the person that can solve your problem. Frankly, getting messages about things that other people can, and should, handle is annoying and doesn’t reflect well on the sender. 
 

Edited June 13 by TechWife