3.1 Generative artificial intelligence
In the previous section, you explored some of the temptations Grace faced when preparing her assignment. There are others, of course, and in this section you will consider one more closely: generative artificial intelligence (GenAI).
A full exploration of what GenAI is and how it works is beyond the scope of this short course. For our purposes here, GenAI is a type of artificial intelligence software which can generate content – including text or images – typically in response to the user’s prompts. Over time, GenAI may bring significant benefits to society; but there is also the potential for negative and damaging results, such as generating fake news and biased or partial outputs.
A further temptation that Grace might face, then, is the potential for using GenAI to tackle her assignments.
Before exploring the potential of GenAI in your studies, you should check your institution’s policies to see whether you are permitted to use GenAI to support your learning. If its use is permitted, check that you fully understand how you are permitted to use it, and how you need to reference your use of GenAI in any work submitted for assessment.
In most institutions, failure to declare using GenAI to produce assessed work would be considered academic misconduct. After all, if someone used GenAI to produce an assessment without declaring its use, they would be deceiving the reader, who would be expecting the work to have been created by the person submitting it. As you will have realised as you have worked through this course, it is important to always be clear ‘what is your own work and what is not’.
Whether your institution permits GenAI or not, it is useful to consider some of the potential concerns with using it to generate academic text:
- Data and information contained within generative AI tools is gathered from a wide range of sources, including ones that are poorly referenced, out of date, or irrelevant. GenAI ‘hallucinates’ (i.e. it makes errors). If you are permitted to use GenAI, it is your responsibility to ensure you are sufficiently skilled in your subject that you recognise when GenAI gets it wrong.
- GenAI produces seemingly original output by learning patterns in its database or from the internet, which predict a sequence of words (in the case of text). It appears to understand the sources it uses and the generated results, but in reality, it does not understand it and the concept of ‘intelligence’ is misleading.
- GenAI tools produce answers based on information generated by humans. You do not know what sources it has used, and these may contain societal biases and stereotypes that, in turn, may be replicated in the generative AI tool’s response.
- Learning to write in an academic style takes time; like any skill, your academic writing will not develop without practice. Overuse of GenAI may reduce your opportunities to practice your writing skills, which you will need throughout your studies, in your workplace and in society in general, where the skills of conducting ethical and accuracy checks on any information you produce or handle will be required.
Interestingly, AI developers around the world have not ignored the need to support the responsible use of GenAI. Techniques such as watermarking AI-generated text are used and continuously improved to keep pace with this rapidly moving field.