+27 21 959 2946library@uwc.ac.za
This work is licensed under CC BY-NC 4.0
Maybe it’s a good thing that the dominance of the college essay has been called into question. The traditional focus on presenting a polished finished project that could just as easily be generated by a chatbot or purchased from a term-paper mill distracts from the dynamic process of thinking through complex questions and weighing possible answers.
Barbara Fister and Alison J. Head
Whether or not you choose to use generative artificial intelligence tools, both teachers and students must comprehend, assess, and generalise their applications. Utilising generative A.I. tools requires you to evaluate whether their use will improve your assignments and assessments through a critical, ethical, and thoughtful lens.
LLMs "save your conversations and can use them as training data," which means that any input or materials uploaded to LMM processors can become part of the model's training set and be shared in the future without attribution. As a result, resources and intellectual property may be used in unexpected ways.
While generative AI tools can help users with such tasks as brainstorming for new ideas, organising existing information, mapping out scholarly discussions, or summarising sources, they are also notorious for not relying fully on factual information or rigorous research strategies. They are known for producing "hallucinations," an AI science term used to describe false information created by the AI system to defend its statements. Oftentimes, these "hallucinations" can be presented in a very confident manner and consist of partially or fully fabricated citations or facts.
Additionally, the information presented by generative AI tools may lack currency as some of the systems do not necessarily have access to the latest information. Rather, they may have been trained on past datasets, thus generating dated representations of current events and the related information landscape.
Generative AI tools have introduced new challenges in academic integrity, particularly related to plagiarism.
Plagiarism is typically defined as presenting someone else's work or ideas as one's own. While a generative AI tool might not qualify as a "someone," using text generated from an AI tool without citing is still considered plagiarism, because the work is still not the researcher's own. Individual policies for using and crediting GAI tools might vary from department or faculty, so looking at the syllabus and having a clear understanding from the professor is important.
A note about plagiarism detection tools:
A number of AI detection tools are currently available to publishers and institutions, but there are concerns about low rates of accuracy and false accusations. Because generative AI tools do not generate large amounts of text word-for-word from existing works, it can be difficult for automated tools to detect plagiarism.
Another area of academic integrity affected by GAI tools is that of false citations. GAI tools such as ChatGPT have been known to generate false citations, and even if the citations represent actual papers, the cited content in ChatGPT might still be inaccurate.
Another potentially significant limitation of AI is the bias that can be embedded in the products it generates. Fed immense amounts of data and text available on the internet, these large language model systems are trained to simply predict the most likely sequence of words in response to a given prompt, and will therefore reflect and perpetuate the biases inherent in the inputted internet information. An additional source of bias lies in the fact that some generative AI tools utilise reinforcement learning with human feedback (RLHF), with the caveat that the human testers used to provide this feedback are themselves non-neutral. Accordingly, generative AI like ChatGPT is documented to have provided output that is socio-politically biased, occasionally even containing sexist, racist, or otherwise offensive information.
(adapted from George Town University Library: https://guides.library.georgetown.edu/ai/ethics)
Links to Further Readings:
Ethical AI for Teaching and Learning
The Ethical Considerations of Artificial Intelligence
Navigating the impacts of generative AI in South Africa: challenges, opportunities and ethics
Practices of AI
With the help of pre-existing artifacts, generative AI is able to produce new, realistic artifacts (at scale) that accurately capture the features of the training set while avoiding duplication. It can generate text, speech, images, video, music, software code, and product designs, among other types of original content. Take a look at the following website by Gartner Experts which looks at what is generative AI and the practices of AI.
| Generative AI Do's | Generative AI Dont's |
 |
 |
| Think critically: Students should think carefully about the output generated by AI tools and take a sensible, skeptical, critical approach. | Rely solely on AI: Students should not blindly rely on AI-generated content and should not use AI tools to create work that is portrayed as their own. |
| Be curious and experiment: Experimenting with generative AI in low-risk situations is the best way to understand it. Many services are freely available, and more are being added on a daily basis. | Cautious: AI shouldn't be used to replace your own knowledge in critical-thinking tasks such as identifying research gaps, developing hypotheses, analyzing data, etc. |
| Use AI to generate ideas or basic content: AI can increase productivity, which can serve as a catalyst for your own creative expression. For example, ask Chat GPT, a generative AI program, to suggest Threads posts that will interest Millennials with children who lead busy lives. | Disregard Ethical Issues: Don't use AI systems that could injure people or reinforce biases. Disregarding ethical issues may result in the implementation of AI systems that have unexpected effects. Prioritizing moral principles and avoiding inventions that might have detrimental effects on society are essential to responsible AI development. |
Evaluation Criteria for Generative AI
To assess the data produced by AI tools like ChatGPT, you may need to take a slightly different approach. You must decide whether the data produced in response to your prompt is reliable enough to serve as a direct source or a foundation for your work while you are a student.
| Currency |
Does the AI tool cite current sources when asked for references or some sample source information, or does the dataset rely on older, publicly available information? How recent is the dataset that your AI tool is trained on? Could you find out? Does this affect the accuracy or applicability of the data? Is it likely that you will need to look for more recent information somewhere else? |
| Relevance |
Is the data produced pertinent to the work you are performing? How well does the answer address the question you posed and the subject you are researching? Do you need to elucidate or expand on the information provided by asking another question? Does the response offer information at a level suitable for the task at hand? Is it sufficiently detailed for scholarly readers? |
| Authority |
Does the generated content have references to sources or supporting evidence? Does it have authority? Can you vouch for the sources? Are the references reliable? Are citations given? Can you confirm that the references are genuine and that they exist? |
| Accuracy | Is the information being provided trustworthy? Is there factual mistakes or details that don't seem to be backed up by any evidence? |
| Purpose |
Does the response seem to be biased in any way? Does it seem as though the text favors one side over the other? |
Further readings:
Citing Generative AI
Before incorporating generative AI into a course or research project, seek guidance from your professor or instructor and the Academic integrity at UWC guide.
If you have permission to utilize generative AI and you incorporate information from a generative AI program, whether through reuse or paraphrasing, and use it to refine your writing, it is essential to acknowledge this practice.
Various citation styles are endeavoring to offer recommendations on whether and how to cite ChatGPT and other generative artificial intelligence technologies. Given the rapid evolution of these technologies, it is expected that the citation guidelines provided below may undergo changes. Therefore, it is advisable to check regularly for any updates.
Authors must clearly acknowledge within the article or book any use of Generative AI tools through a statement which includes: the full name of the tool used (with version number), how it was used, and the reason for use. For article submissions, this statement must be included in the Methods or Acknowledgments section.... ....This level of transparency ensures that editors can assess whether Generative AI tools have been used and whether they have been used responsibly.
Taylor and Francis Guidelines to Authors
If you use generative AI in the development of ideas and concepts for your assignment, you should acknowledge in your bibliography how you used the tool, even if you don’t include any AI generated content in your work. Provide a description of the AI tool used, how the information was generated, including the prompts you used, and the date accessed. You could add this acknowledgement as a footnote or at the end of your reference list but you should check with the course coordinator to clarify.
Format:
I acknowledge the use of [insert AI system(s) with link] to [how the AI was used]. The prompts used include [list of prompts].
Example:
I acknowledge the use of ChatGPT [https://chat.openai.com/] to generate ideas and material for background research and project planning in the drafting of this assignment. The following prompts were entered into ChatGPT on 30th March 2023:
|
|
This blog post offers an in-depth guide on citing ChatGPT using the APA referencing style. |
|
|
This site offers a comprehensive guide on adhering to the MLA referencing style when citing ChatGPT |
|
|
The Q&A section of the Chicago Manual Style Online provides guidance on citing generative AI. |
|
|
This site provides a comprehensive guide on how to cite generative AI using different referencing styles including the Harvard referencing style. |
Understanding Copyright
Generative AI tools, such as ChatGPT, commonly give rise to two distinct copyright concerns. The first issue, known as the "input question," revolves around the utilization of copyrighted materials during the training process. Frequently, these materials are sourced from the internet without obtaining permission from the original creators. Legal action against the developers of AI tools may result from claims made by some copyright holders that using their content in this way constitutes a copyright infringement.
In response, AI developers contend that the fair use doctrine within copyright law allows for such use, a viewpoint generally supported by legal experts. The second set of questions focuses on the "outputs" of AI, exploring whether the creations generated by these tools violate the copyrights of their source materials and qualify for copyright protection as distinct works of authorship (Lucchi, 2023).
University of the Western Cape,
Robert Sobukwe Road,
Bellville,
7535
Tel: 021 959 2946
