Skip to Main Content

AI for Researchers

Overview

This resource has been created to provide guidance to all people engaged in research at UniSA, including academic staff, adjuncts and students (research degree, post-graduate, honours and undergraduate).

Artificial Intelligence (AI) refers broadly to any technology that enables machines to simulate human intelligence and perform tasks like sensing, comprehending, acting and learning. AI encompasses a wide range of tools and techniques (taken from the article Alistair linked to in this channel. (source: https://www.linkedin.com/pulse/understanding-difference-between-ai-ml-genai-fullcreek-qwzzc/). Non-generative AI is not new – in fact you’ve already been using it! For example, algorithms are AI, so is predictive text and things like SpellCheck and Grammarly. Speed cameras that can ‘read’ number plates use AI. 

Generative Artificial Intelligence (GenAI) is the branch of artificial intelligence systems capable of producing human-like responses to specified user input, usually in the form of text prompts. GenAI systems encode information present in its training data and generate probable responses based on specific user inputs.

GenAI brings both opportunity and risk when used in research. Improved efficiencies and productivity are clear advantages to using these models, for example in solving data-driven problems faster or perceived enhanced readability of research outputs. One key risk when using GenAI in research is the potential for breaches in research integrity including matters related to confidentiality, intellectual property and authorship.

The University expects that all researchers uphold the highest standards of research integrity by abiding to the principles of responsible research as outlined in the Australian Code for the Responsible Conduct of Research (Code) and the University’s Research Integrity Policy. When using GenAI in research it is the researcher’s responsibility to ensure the principles in the Code and the University’s policy are adhered to; non-compliance may result in a breach of research integrity.