Ulster University

AI Applications for Teaching and Research

 

This site is a repository for case studies of AI use in teaching and research, to provide a platform for specific examples of responsible AI use and to promote such use more broadly. It is open to use cases from Ulster University, Queen’s University Belfast, and potentially other institutions.

This cross-institution availability will allow for a more comprehensive understanding of AI use and better understanding of whether different institutional policies affect the distribution of AI use.

The more use cases that are received, the greater the potential for understanding different possible applications of AI for teaching and research.

Suggest a new use case

USE CASES
Teaching Application(s)Skills neededBenefitsOther links
  • Communication
  • Research design
  • Convai (LLM interface engine for 3D avatars)
 
  • Generative Large Language Model (LLM)
Prior skills: Basic to intermediate level of familiarity with Unity 3D. Some upskilling required with C# coding. Short term: It decreases the time to design NPCs (non-playable characters) and incorporate LLMs in them to reply faster and with accuracy. Long term: After a series of training rounds, the NPCs will be capable of providing training for conflict resolution practitioners through the simulation of conflict scenarios.  AlvidrezSalvadorCommunication and MediaAHSSUlster University
  • Art
  • Design
  • Research design
 
  • Text to image: Midjourney
  • Patterned AI
  • Stable Diffusion XL
Prior skills: Familiarity with design software packages; Text to image AI applications. Upskilling required with Python Coding with understanding of HPC and access to GPU. Upskilling with Stable Diffusion XLShort Term: Significantly speeding up the design research process to create visuals and photorealistic images enabling early cost-effective design decisions to be made without costly resources. Long Term: Potential to contribute to addressing sustainability in fashion industry — Increased accuracy in trends prediction, reducing deadstock and overstock and ultimately landfill. CoulterJanetBelfast School of ArtAHSSUlster University
  • Linguistics
  • Language acquisition
  • Analysis
 
  • Automatic speech recognition (ASR)
Transcription with corpus formatting, Python  GerardJulianaCommunication and MediaAHSSUlster University
  • Linguistics
  • Law
  • Data retrieval
  • Thematic analysis
 
  • Vector embedding
  • Generative Large Language Model (LLM)
Python, XML, web designRapid data classification and analysis MacleodMorganCommunication and MediaAHSSUlster University
  • Psychology
  • Public Health
  • Epidemiology
  • Survey research
  • Data harmonization
  • Scale development
 
  • Natural language processing
  • SBERT transformer neural network
Prior skills: Basic familiarity with emerging AI applications Upskilling required: engagement with capabilities of NLP at the conceptual level. Additional appreciation and skills with general machine learning methods. Short term: Assists with identifying and providing a consistent and replicable indicator of semantic similarity between survey variables from different sources. Long term: Will speed up and increase the replicability of data discoverability and harmonization processes in survey-based research.  McElroyEoinPsychologyLife and Health SciencesUlster University