Interim Guidance on ChatGPT and other AI Generative Resources

Interim Guidance on ChatGPT and other AI Generative Resources

by Stuart Jonker -
Number of replies: 0

At the request of David Hohne, our Academic Dean, here is Moore College's interim guidance on using ChatGPT and related generative technologies in your studies at Moore. Note that this is interim guidance and is subject to change. Please watch for updates, which may include significant changes.

 

 

Interim Guidance on ChatGPT and other AI Generative Resources 

With the advent of ChatGPT, students and teachers now have another electronic tool to use in their research and learning. Nearly all of us use Google for quick searches and Wikipedia for quick introductions and summaries, and all of us should use Grammarly Premium to assist in any important writing. Now, ChatGPT has arrived, and it is a potentially useful tool, though not always reliable. 

The following is interim guidance on the acceptable use and referencing of ChatGPT and similar generative tools.  

Note: These guidelines are interim only and may change in the future. We will notify you of any changes in policy or advice; you are responsible to read all such emails and other official notifications. 

 

Guidance on the use of ChatGPT and other AI-generative technologies

Students may

  •  Use ChatGPT and similar sources for initial research, overviews, explanations, and other basic research. 
  • Use it for initial and provisional answers. 
  • Use it as you would resources such as Wikipedia, or the popular sites that Google may send you to. 
  • You may summarise, paraphrase or quote such material, though this is discouraged for many reasons, and especially for reliability.  If you do utilise such material, you must acknowledge it with a footnote and an acknowledgement declaration, as outlined below. 

For these and other uses in your academic work, regard ChatGPT and other generative technologies as you would Wikipedia: It may give some reliable information or references, but it is not consistently reliable. Note College’s serious cautions and reservations below. 

 

Students must

  • Clearly and fully reference any information, ideas, or text generated by such technologies. This includes outlines, ideas, points within an answer, and any other information or parts of your submitted answer.  
  • State that you have used ChatGPT and other generative technologies even if you have not used them in your final written or oral submission.  

Referencing and acknowledgement guidelines follow. 

 

Students must not

  • Violate the spirit of the MTC Academic Integrity policy. 
  • Present any material from ChatGPT or similar sources in a way that leaves the impression it was your original work, thought, or conclusions. 

 

Referencing AI-generated technologies and information 

 

If you use an AI-generative technology such as ChatGPT at any stage of your research, you need to declare that. If you use it as a source, you need to footnote that. 

 

 

Reference it as you would other ephemeral sources, such as email and personal correspondence: 

 

Referencing the Use of Al Technologies 

 

All direct quotes, summaries, and paraphrases of text Al sources must be referenced in a footnote. Students are also required to add a statement outlining their use of Al technologies if they use them in an assignment. Examples follow.  

 

SBL Referencing 

 

Footnotes:     

 

(number) Originator of the communication, medium, Day, Month, Year  

 

1 Open Al's ChatGPT Al language model, response to question from the author, 17 February 2023. 

 

Note: Use this full form of the footnote each time you need a footnote. As yet, there is no abbreviated form for second and subsequent footnote. 

 

Bibliography:  

An informal source; do not reference, but you must declare its use (see template below). 

 

 

 

Declare your use of it with this declaration: 

 

Declaration of Al technology use 

 

If you use Al, you must include a declaration outlining your use of it and the programs used. This will help establish your academic integrity in this work. 

 

Declarations should follow the Bibliography of Sources Used. 

 

 

Declaration of use (When you use AI-generated information in the research process or for background only, and it does not form part of your final submission) 

 

Template

I acknowledge the use of <insert Al system and link to the system> to <insert purpose of Al use> in <insert which stage of the assessment>.  

 

Examples

I acknowledge the use of ChatGPT, https://openai.com/blog/chatgpt/ to assist with the structuring of this essay in the drafting process.  

 

I acknowledge the use of Elicit https://elicit.org/ in collating resources for use during my initial research. 

 

 

Declaration of modified use (When you use and quote, paraphrase, summarise, or otherwise rely upon AI-generated information in your submission) 

 

Template

I acknowledge the use of <insert Al system and link to the system> to generate <insert materials> which were [modified and] used <insert which stage of the assignment>. 

 

Examples

I acknowledge the use of Synthesia https://www.synthesia.io/ to generate a short video about using Al, which was then modified and used in my final oral presentation. 

 

 

(Adapted from Summary of Interim Al Technologies Referencing - Update February 2023 by Tabor Institute of Higher Education) 

 

 

Limitations and possibilities 

Possibilities

There are positive and wise uses of these technologies. Some students have told me they use ChatGPT during lectures to answer quick questions. As they report, “It pinballs us back into the lecture,” and saves yet another clarifying question. And there are other sources coming out that will list or suggest academic bibliographies on various topics. (They are not yet very good.) These are early days. The technology will develop, as will our imagination on how to use it. (Note: As you develop other positive uses,
Gordon Cain would love to hear about them.) 

 

Limitations

As you probably know, ChatGPT and similar technologies offer new, machine-generated, unique answers each time they are prompted. These programs are created to produce human-like texts, not accurate or true texts. Their answers are based on many terabytes of randomly collected information. These include academic texts of varying quality and many other texts of all types. From these, it generates answers that conform to a pattern (eg, a three-point essay) with no regard for the quality of the information

AI generators like ChatGPT are not good starting points for academic research. Sometimes its answers are correct, but often, these generative technologies provide conveniently quick answers that are incomplete or wrong. This is because every answer is unique. It also generates ‘hallucinatory’ references that appear reliable but do not exist. Thus, if you are new to a topic, you have no way of knowing if an answer is reliable, accurate, or complete. 

It is far better to start with a few reliable overviews such as dictionary articles, recent commentaries, or survey chapters and articles. 

In summary, ChatGPT and its cousins are convenient but not reliable sources. Moore College strongly discourages students from using them for academic research. Alternatives for informal research are: 

  • Google search – Possibly more reliable, if you know how to check the quality of your sources with the CRAP test
  • Google Scholar – Academic sources. The ATLA database is usually preferable for your Biblical and Christian studies. 
  • Wikipedia – Often a general or reliable overview and may lead to good sources. It is usually somewhat balanced because of its open editing format. As with any other source, you still need expert knowledge to discern quality or bias. 
  • Grammarly – Use it and not ChatGPT to improve your formal, academic writing style. It gives better results, is simpler to use well, and College pays for it. 

 

An encouragement

Remember, one of your goals at College is to become a wise and skilled minister of the word and pastoral thinker. These tools mitigate against that by providing prefabricated answers that may or may not accord with the Lord, his word, and his wisdom. 

 

Interim guidance, ver 1.0, 2023-04-19