Academic misconduct. Plagiarism, authenticity and integrity.

The most fundamental concern is the potential for automated cheating and using AI to create in whole or part, essays, reports, images, or academic answers without properly attributing the source. This is obviously a serious issue as it is contrary to the fundamental ethos of learning and assessment.

Bias and Discrimination

The data that is used to train the AI model directly affects the data that is generated from it. Therefore inherent bias is found in the major models, which can lead to the AI generating output that is biased, discriminatory and reliant on stereotypes. 

Intellectual property, ownership and copyright

Scraping vast amounts of data from the internet to train the AI models, has led to many contentious arguments as to the both the ethics of and the legality of training AI on data without the consent of copyright owners.

A second issue is who owns the data that has been generated by the AI. Human prompts are only part of the story when it comes to the origin of the data.

Often students' work is being used to train the AI models and institutions must make sure they have permission before allowing this to happen.

Creative critical thinking and originality of academic development

Overreliance on AI to define the boundaries of a subject and obtain requisite data sources, can lead the student to disengage from the essential subject knowledge and reading. Generative AI is only as good as the training data, in terms of scope, age, coverage, and reliability. 

All areas of academic endeavour develops from what has been done before. The phrase “standing on the shoulders of giants” (Newton, 1675), means that your original work can move on from the agreed academic accepted understanding that has gone before. This development means you don’t have to redo the earlier work, but the development of theories, finding solutions and answers to problems that are posited, are your original work. 

Generative AI merely generates answers based on existing data without having the capacity to create those original approaches that true academic development requires.

Referencing using the APA for SHU validated students

AI generated data isn’t produced by a human author and as such can’t be referenced in the usual way. It is unpublished data, that cannot be ‘found’ on the internet, like other data can.

Referencing format

Author: Usually this will be the company responsible for producing the AI Google

Date: The year the version used was produced e.g. (2025)

Title: The name of the AI model and the version in round brackets e.g. Gemini (2.5 Flash)

Description: Description of the AI model in square brackets e.g. [Large language model]

Source/URL: The link to the AI interface e.g. https://gemini.google.com

Example citation - (Google, 2025)

Example Reference - Google. (2025). Gemini (2.5 Flash) [Large language model]. https://gemini.google.com