Gen AI and Ethical Considerations

Limitations and biases

GenAI cannot think critically. It can also give out false information, which might be perceived as sexist or racist for example. GenAI may produce biased or misrepresented outputs, reinforcing stereotypes or discriminating against certain groups. Additionally, GenAI may not “understand” context or nuance, leading to misinterpretations or errors in its output. Furthermore, its reliance on algorithms and patterns can result in a lack of creativity or originality, limiting its ability to generate truly innovative content.

Ethical considerations

There are a range of ethical issues associated with GenAI, particularly related to how the models have been developed including the information used to train them. For example, GenAI has been trained on copyrighted materials without permission from the original creators and in the use of photographs, consent was not sought from subjects. Some GenAI tools utilise user input to further train and refine their models and collect user data, raising concerns about transparency and data privacy, as well as issues around users sharing information that does not belong to them, such as company data.

There are inherent biases in the GenAI outputs due to the lack of diversity in the data used to train the models. Ethical concerns are also raised around the use of AI to make decisions in professions such as healthcare as well as its potential to create illegal content such as child pornography. There are other human rights and sustainability issues associated with GenAI, including use of low paid workers to “clean” data, equity of access for GenAI tools only available through subscription and the amount of water used by GenAI technologies.

(Sources: AI's excessive water consumption threatens to drown out its environmental contributions, https://theconversation.com/ais-excessive-water-consumption-threatens-to-drown-out-its-environmental-contributions-225854 ; AI tools: Issues and Considerations - University of Newcastle Libguides, https://libguides.newcastle.edu.au/AI-tools/issues)

Copyright considerations

GenAI information is usually considered public domain, this means that it can be used freely, within the academic integrity guidelines of the university context.  However, as considered in the Ethical Considerations above, there is no way to verify if the content generated by AI has been taken from copyright protected materials without permission.

You must never upload copyright protected material such as published journal articles, book chapters, or UC publications such as lecture notes, to GenAI tools as they may be illegally stored by the system and used to train the AI.

Open and transparent use

At UC, currently GenAI can only be used with permission from your unit convener and as part of the assessment instructions. To use GenAI appropriately, it is important to acknowledge how it was used it, e.g. to brainstorm your initial ideas or to produce an essay outline. The use of GenAI must be referenced accurately following the UC Referencing Guide.

Thinking critically when using GenAI

When using GenAI, it is important for students to analyse the information provided, evaluate perspectives and develop logical arguments. It is important not to accept what GenAI offers without reading, understanding and evaluating it first. It is important to identify any bias and limitations in the output. Students should always question the validity of GenAI output and seek out authoritative sources to validate the GenAI outputs.