A whole new ballgame in cheating. Introducing ChatGPT

[Kron3007]

Here's another article from the NYT:

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach
With the rise of the popular new chatbot ChatGPT, colleges are restructuring some courses and taking preventive measures.

https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html

ChatGPT weighs in:

"As the use of AI-powered language models such as ChatGPT becomes more prevalent, professors and educators are expressing concerns about the potential for students to use these tools to cheat on assignments and exams.

One of the main concerns is that students may use ChatGPT to generate answers to homework and test questions, potentially allowing them to pass off someone else's work as their own. Additionally, students may use the model to generate entire papers or essays, bypassing the need to do any original research or writing.

To address these concerns, some professors are taking steps to detect and prevent the use of ChatGPT and other AI language models by students. One approach is to use plagiarism detection software, which can flag papers or essays that have been generated by a machine. Another approach is to require students to submit their work in a format that is not easily readable by a machine, such as handwriting or a PDF.

Additionally, some professors are incorporating more open-ended and creative assignments, which are less amenable to cheating with AI models. Some are also incorporating more group work and discussions, which can make it more difficult for students to pass off someone else's work as their own.

It is important to note that while ChatGPT and other AI language models can be used for academic dishonesty, they can also be used for legitimate educational purposes, such as generating writing prompts or providing feedback on essays. As such, it is important to strike a balance between preventing cheating and promoting the use of these technologies for educational advancement.

Overall, Professors and educators are concerned about the potential for students to cheat using language models like ChatGPT, and are taking steps to address this concern. However, it's important to remember that these models can also be used for legitimate educational purposes."

[marshwiggle]

Here's another article from the NYT:

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach
With the rise of the popular new chatbot ChatGPT, colleges are restructuring some courses and taking preventive measures.

https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html

ChatGPT weighs in:

"As the use of AI-powered language models such as ChatGPT becomes more prevalent, professors and educators are expressing concerns about the potential for students to use these tools to cheat on assignments and exams.

One of the main concerns is that students may use ChatGPT to generate answers to homework and test questions, potentially allowing them to pass off someone else's work as their own. Additionally, students may use the model to generate entire papers or essays, bypassing the need to do any original research or writing.

To address these concerns, some professors are taking steps to detect and prevent the use of ChatGPT and other AI language models by students. One approach is to use plagiarism detection software, which can flag papers or essays that have been generated by a machine. Another approach is to require students to submit their work in a format that is not easily readable by a machine, such as handwriting or a PDF.

Additionally, some professors are incorporating more open-ended and creative assignments, which are less amenable to cheating with AI models. Some are also incorporating more group work and discussions, which can make it more difficult for students to pass off someone else's work as their own.

It is important to note that while ChatGPT and other AI language models can be used for academic dishonesty, they can also be used for legitimate educational purposes, such as generating writing prompts or providing feedback on essays. As such, it is important to strike a balance between preventing cheating and promoting the use of these technologies for educational advancement.

Overall, Professors and educators are concerned about the potential for students to cheat using language models like ChatGPT, and are taking steps to address this concern. However, it's important to remember that these models can also be used for legitimate educational purposes."

I think I'm already starting to get used to the "non-answer" answer it generates; no real insight into how to detect it, just an assertion that it has uses.

[Kron3007]

Here's another article from the NYT:

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach
With the rise of the popular new chatbot ChatGPT, colleges are restructuring some courses and taking preventive measures.

https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html

ChatGPT weighs in:

"As the use of AI-powered language models such as ChatGPT becomes more prevalent, professors and educators are expressing concerns about the potential for students to use these tools to cheat on assignments and exams.

One of the main concerns is that students may use ChatGPT to generate answers to homework and test questions, potentially allowing them to pass off someone else's work as their own. Additionally, students may use the model to generate entire papers or essays, bypassing the need to do any original research or writing.

To address these concerns, some professors are taking steps to detect and prevent the use of ChatGPT and other AI language models by students. One approach is to use plagiarism detection software, which can flag papers or essays that have been generated by a machine. Another approach is to require students to submit their work in a format that is not easily readable by a machine, such as handwriting or a PDF.

Additionally, some professors are incorporating more open-ended and creative assignments, which are less amenable to cheating with AI models. Some are also incorporating more group work and discussions, which can make it more difficult for students to pass off someone else's work as their own.

It is important to note that while ChatGPT and other AI language models can be used for academic dishonesty, they can also be used for legitimate educational purposes, such as generating writing prompts or providing feedback on essays. As such, it is important to strike a balance between preventing cheating and promoting the use of these technologies for educational advancement.

Overall, Professors and educators are concerned about the potential for students to cheat using language models like ChatGPT, and are taking steps to address this concern. However, it's important to remember that these models can also be used for legitimate educational purposes."

I think I'm already starting to get used to the "non-answer" answer it generates; no real insight into how to detect it, just an assertion that it has uses.

Yeah, can't say it is the most insightful article on the subject, but pretty good for the 30 seconds it took.  Grammatically, it has some of my students beat! 

[jimbogumbo]

https://www.thewrap.com/cnet-removes-staff-byline-ai-generated-stories/

[dismalist]

From the article:

The "CNET Money Staff" has generated more than 70 AI-generated financial articles to date, with such titles such as "Should You Break an Early CD for a Better Rate?"

If I knew how to make money in the markets, why should I tell you?

[secundem_artem]

I just tried it to answer some of the usual questions I may ask students.  Not a lot of depth but a passable response for a FR or SO.  I asked another question requesting a specific length and a reference list and got back a very short paper with 3 references.

As to writing for publication, I asked it to answer the question we asked on my most recent piece of research.  Got back a very poor response.

I think it's going to make us dig a lot deeper in setting up questions - which is going to play hell with student grades and any subsequent teaching evaluations.  Basically, courses are going to have to get a LOT more difficult.

Best case scenario is that artificial intelligence will never be able to beat out actual stupidity.

[marshwiggle]

I just tried it to answer some of the usual questions I may ask students.  Not a lot of depth but a passable response for a FR or SO.  I asked another question requesting a specific length and a reference list and got back a very short paper with 3 references.

As to writing for publication, I asked it to answer the question we asked on my most recent piece of research.  Got back a very poor response.

I think it's going to make us dig a lot deeper in setting up questions - which is going to play hell with student grades and any subsequent teaching evaluations.  Basically, courses are going to have to get a LOT more difficult.

It would be terribly ironic if AI wound up being the thing that resulted in raising the standards for higher education so that it ceased being the destination for all kinds of students who aren't prepared for it.

ETA: Will the former "gentleman's C" become the "AI's C"?

Best case scenario is that artificial intelligence will never be able to beat out actual stupidity.

No danger; real stupidity requires real, conscious immersion in a universe of "alternative facts". An AI may be able to get things wrong, but it won't have the total commitment required for true stupidity.

[Kron3007]

I just tried it to answer some of the usual questions I may ask students.  Not a lot of depth but a passable response for a FR or SO.  I asked another question requesting a specific length and a reference list and got back a very short paper with 3 references.

As to writing for publication, I asked it to answer the question we asked on my most recent piece of research.  Got back a very poor response.

I think it's going to make us dig a lot deeper in setting up questions - which is going to play hell with student grades and any subsequent teaching evaluations.  Basically, courses are going to have to get a LOT more difficult.

Best case scenario is that artificial intelligence will never be able to beat out actual stupidity.

Yes, but there are many ways to use it.  I personally see it providing value as a template, that could then be fleshed out to save some time. In its current form, simply using the output as is would be marginal at best, but could be beefed up and get a decent grade I think.

Is this really cheating?  Maybe, but in the future, a lot of writing will likely be computer assisted and this will expected. I'm sure when spell check came out people may have viewed it as cheating, but now it is standard.  Likewise with grammarly etc , is it cheating to have a computer help you with your grammar? 

Interesting times....

[Larimar]

In my class, use of AI for writing is cheating. I teach first year comp. The purpose of the class is to teach them how to write a coherent and persuasive argument. Using AI to write their papers would not be doing their own work.

Would it be a different story for other classes and other fields? Perhaps. However, AI should not replace language skills. Where are we without language and communication? As for grammar and spell checkers and citation generators, they are helpful tools, and I allow students to use them, but I also explain to my students how these functions work and show in class that they are not infallible. There's no real substitute for doing these things oneself.


Larimar

[apl68]

Here's another article from the NYT:

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach
With the rise of the popular new chatbot ChatGPT, colleges are restructuring some courses and taking preventive measures.

https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html

ChatGPT weighs in:

"As the use of AI-powered language models such as ChatGPT becomes more prevalent, professors and educators are expressing concerns about the potential for students to use these tools to cheat on assignments and exams.

One of the main concerns is that students may use ChatGPT to generate answers to homework and test questions, potentially allowing them to pass off someone else's work as their own. Additionally, students may use the model to generate entire papers or essays, bypassing the need to do any original research or writing.

To address these concerns, some professors are taking steps to detect and prevent the use of ChatGPT and other AI language models by students. One approach is to use plagiarism detection software, which can flag papers or essays that have been generated by a machine. Another approach is to require students to submit their work in a format that is not easily readable by a machine, such as handwriting or a PDF.

Additionally, some professors are incorporating more open-ended and creative assignments, which are less amenable to cheating with AI models. Some are also incorporating more group work and discussions, which can make it more difficult for students to pass off someone else's work as their own.

It is important to note that while ChatGPT and other AI language models can be used for academic dishonesty, they can also be used for legitimate educational purposes, such as generating writing prompts or providing feedback on essays. As such, it is important to strike a balance between preventing cheating and promoting the use of these technologies for educational advancement.

Overall, Professors and educators are concerned about the potential for students to cheat using language models like ChatGPT, and are taking steps to address this concern. However, it's important to remember that these models can also be used for legitimate educational purposes."

I think I'm already starting to get used to the "non-answer" answer it generates; no real insight into how to detect it, just an assertion that it has uses.

Yeah, can't say it is the most insightful article on the subject, but pretty good for the 30 seconds it took.  Grammatically, it has some of my students beat!

This may be ChatGPT's biggest use right here--generating canned institutional responses!  Most of them have looked like the products of bots for years now anyway.

[apl68]

In my class, use of AI for writing is cheating. I teach first year comp. The purpose of the class is to teach them how to write a coherent and persuasive argument. Using AI to write their papers would not be doing their own work.

Would it be a different story for other classes and other fields? Perhaps. However, AI should not replace language skills. Where are we without language and communication? As for grammar and spell checkers and citation generators, they are helpful tools, and I allow students to use them, but I also explain to my students how these functions work and show in class that they are not infallible. There's no real substitute for doing these things oneself.


Larimar

This makes me think about mathematics and calculators.  Sure, calculators can give quick, simple answers so that one never needs to know things like fractions, multiplication tables, etc.  But if one never studies these sorts of basics, how is one ever going to develop any knowledge of mathematics?  An AI might let you cheat your way to a passing grade in a low-level comp class, but students who do this are cheating themselves right out of any learning they might have done.

[Larimar]

In my class, use of AI for writing is cheating. I teach first year comp. The purpose of the class is to teach them how to write a coherent and persuasive argument. Using AI to write their papers would not be doing their own work.

Would it be a different story for other classes and other fields? Perhaps. However, AI should not replace language skills. Where are we without language and communication? As for grammar and spell checkers and citation generators, they are helpful tools, and I allow students to use them, but I also explain to my students how these functions work and show in class that they are not infallible. There's no real substitute for doing these things oneself.


Larimar

This makes me think about mathematics and calculators.  Sure, calculators can give quick, simple answers so that one never needs to know things like fractions, multiplication tables, etc.  But if one never studies these sorts of basics, how is one ever going to develop any knowledge of mathematics?  An AI might let you cheat your way to a passing grade in a low-level comp class, but students who do this are cheating themselves right out of any learning they might have done.

Thank you, and yes, I completely agree with your mathematics analogy.


Larimar

[Kron3007]

In my class, use of AI for writing is cheating. I teach first year comp. The purpose of the class is to teach them how to write a coherent and persuasive argument. Using AI to write their papers would not be doing their own work.

Would it be a different story for other classes and other fields? Perhaps. However, AI should not replace language skills. Where are we without language and communication? As for grammar and spell checkers and citation generators, they are helpful tools, and I allow students to use them, but I also explain to my students how these functions work and show in class that they are not infallible. There's no real substitute for doing these things oneself.


Larimar

This makes me think about mathematics and calculators.  Sure, calculators can give quick, simple answers so that one never needs to know things like fractions, multiplication tables, etc.  But if one never studies these sorts of basics, how is one ever going to develop any knowledge of mathematics?  An AI might let you cheat your way to a passing grade in a low-level comp class, but students who do this are cheating themselves right out of any learning they might have done.

Thank you, and yes, I completely agree with your mathematics analogy.


Larimar

Yes, I think for a comp class that makes perfect sense.  I do wonder why you accept grammar checkers in this type of class, but I take your point.

Regarding calculators, it is definitely a good example and shows how we have accepted the use of calculators and now see using them as a given (not during early years to learn the very basics of course).  I can't imagine doing trig without one to be honest, but I'm sure there was a time where this would have been blasphemy.  Likewise with statistics, science would slow to a crawl without computers doing the heavy lifting...     

[Wahoo Redux]

I do wonder why you accept grammar checkers

Well, there is no real way to detect grammar checkers, especially since grammar check is now a basic function in MSWord.  And somehow they still get a lot of grammar wrong.

When I taught writing I actually had a little routine where we went through the different cues in Word ("You see that little red squiggly line..." etc.).  I told students to let the program help teach them grammar and syntax---it's helped me with mine.  After all, most grammar experts think that learning grammar in context is about the only way to do it.  The Word grammar-check is pretty good at doing this, and students still have to understand the rules because Word is not infallible and a bit anal-retentive.

And I think this is the way we can use AI.  They can be personal tutors if used correctly. These programs can actually help us learn stuff. 

[Puget]

And I think this is the way we can use AI.  They can be personal tutors if used correctly. These programs can actually help us learn stuff.

I do think there is huge potential for using it as an interactive tutor, e.g., students could put in their understanding of some topic and ask it to identify the parts they got wrong and correct them. (Automated *grading* is another obvious use!).

Another good use could be making study mnemonics -- some of the best things I've seen posted are were people have asked it to write or song or limerick about an academic topic, or explain it in the style of e.g., a pirate, a drunk frat bro, etc. They are really funny and surely more memorable than dry prose. I may suggest this to my seminar students this semester (as long as they carefully fact-check the results), maybe even encourage them to share them with the class.