Using generative AI at university

Academic integrity and AI

Academic integrity means being honest about how you produced your work. At La Trobe, the AI Assessment Scale sets clear expectations for each assessment task — so the starting point is always checking what level of AI use is expected for your specific assessment, not whether AI use is permitted in general.

Whilst AI tools can be helpful in supporting your learning, you remain responsible for verifying and validating all information submitted as part of your assessment. Failure to do so puts you at risk of academic misconduct, which may impact your progress. 

If you're unsure what's expected, check your Assignment Task Sheet or ask your Subject Coordinator.

Key questions to ask yourself before using AI:

• What level of AI use does my assessment allow?
• Am I using this tool to support my thinking, or to replace it?
• If I'm using AI as a research aid, am I locating, reading, and citing the original sources — not just accepting AI-generated summaries?
• Have I checked the content for accuracy, bias, and whether it genuinely supports my argument?
• Have I acknowledged my AI use in line with my subject's instructions?
• Am I uploading anyone else's work or confidential information into this tool?
   

Privacy and online safety

Privacy and online safety is important to consider when using generative AI tools as with any other online activity. You should also be cautious of putting personal data and intellectual property into generative AI tools because many of these tools will collect and store information you provide. This data could be used to continue training the platform or sold to third parties.  
 

Accuracy and reliability

Generative AI tools attempt to provide the most probable response based on their training data, which means outputs can be inaccurate, incomplete, or misleading even when they appear authoritative. A common issue in academic settings is hallucination — where AI generates false or unverifiable citations. However, a more subtle and increasingly common risk is when AI accurately cites a real, published source but misrepresents what that source actually says. Always locate and read original sources rather than relying on AI-generated summaries. The SIFT method is a useful strategy for evaluating the credibility and accuracy of AI-generated content.