Today, the OPSITAra conference begins at Ara Institute of Canterbury. The conference brings researchers from Ara, Otago Polytechnic, and Southern Institute of Technology predominantly and with other ITPs, to together to share their work.
The conference begins with a Powhiri (Māori welcome) followed by catch up networking with morning tea :) Darren Mitchel, the Ara CE and currently also for Rohe 4 for Te Pūkenga, welcomes all conference participants. Dr; Jamie Smiler, research / rangahau director for Te Pūkenga reiterated the welcome and updated on the activities across the week with similar conferences at Unitec, the Pacific forum and a virtual conference yesterday. He encouraged research to be relevant to our communities and that research should not just live on a bookshelf, but be applied to supporting and improving the lives of many.
Scott Klenner, research manager for Otago Polytechnic and Ara introduces the first conference keynote Dr. Waikaremoana Waitoki, who discusses 'growing researcher practice and praxis' - He piko he taniwha, he taniwha rau -on every bend a chief. Began with a tribute to Professor Angus Macfarlane, who passed on last week and who shoulder tapped Moana to present this keynote. Moana summarised her whakapapa and her research objectives. She wanted to ensure the kaupapa of supporting research impact was followed through. What does it mean to conduct research that critically engages with the lived realities of communities?How can researchers ensure their work meaningfully contributes of societal transformation? and also covered some of the critical indigenous research methodologies emphasise relationality, reciprocity, cultural integrity and the co-creation of solutions, rooted in community values. Anchored work in 'the politics of distrction' - especially in the current political climate. Referred to the work of Smith (1999) - the indigenous research agenda - what are the factors that support self-determination?
Posed the challenge to researchers to think through what is indigenous Māori? what is mātautanga Māori and how these inform ethics, funding proposals, student research, responses to complaints from lobby groups etc. Discussed racism and what happens when different groups of people, think of others as different and classify them. Used the example of the government's cutting humanities and social science research through the Marsden fund, to devalue Māori research.
Shared a recent Marsden funded project on mātauranga Māori in psychology. Explained what was Mauri - as the energy that influences the emotional and spiritual aspects of life. Developed the concept of Mauri Ora Tai Pari - a awareness of how various influences from other people, environment etc. activates and arouses us and these cycles are now applied to mental health, alcohol/drug addictions etc. to inform care and support. The Mauri ora tako is a consolidation of the many frameworks to support wellbeing (Waitoki & McLauchlan, 2020) and applied to the work of psychologists. with a inter-relational and holistic approach. Application of the approach during the Waakari (White Island) eruption.
Then shared several other projects including raranga, raranga taku takapau: Hapū ora for tamariki and examples of implications. Recommended 3 books to follow up and the people we should be reading. Shared that Mauri books also available online and for tamariki - children. Shared current project and work. Encouraged researchers to ensure Te Tiriti is honored and the importance of making a submission to the current treaty principles bill submission.
After the keynote, I chair the first sessions in the Education 1 stream.
First up, John G. Mumford from Southern Institute of Technology on 'machine learning rules interpretability: critical perspectives for postgraduate IT students. Shared the research scope, background, focus of the presentation, implications and conclusion. Study explored the application of critical thinking to understand how cognitive biases play a role in hindering the interpretability of machine learning rules, some of the ways debiasing strategies can be used, the existence of the interpretability / accuracy trade off, and shed light on the black box of AI. key themes brought together through a scoping literature (2020 onwards) review. Key cognitive biases - awareness of selected cognitive biases, awareness of interpretability/accuracy trade off, debiasing strategies, clarifying ML rules and enhanced interpretability and the ML black box. Cognitive bias through primacy effect, information bias, and misinpretation, ambiguity aversion and insensitivity to sample size. Implications include helping students to navigate the black box, apply logic and statistical knowledge to Ml roles, navigate teachers and students biases and help learning techniques for debiasing.
Then Robert Nelson, Sam Mann and Ruth Meyers from Otago Polytechnic (OP) present in 'raging with the machine: Collaborating with AI in learning.' Sam presented on a chapter from Robert's Doctorate in Professional Practice thesis, studying collaborative project-based learning. Context is a one semester course where a project is formulated and 'solved'. The actual way the brief is formed, impinges on the context, collaborations and authenticity of the project. AI now pervasive and authenticity requires that AI is deployed. AI can be part of how the work is undertaken but AI can also be a collaborator with more that one AI possible in the project. Another option is to get AI to do the work and the human is the 'conductor' or the human could just be the observer of AI as they complete the project. Experimented with a project and found that the job tasks have accelerated! - from weeks to 2 hours. Raises the challenge of why we take a project across a whole semester. With AI, several projects can be undertaken in a semester! and is this desirable?
Third presentation is with Bruno Balducci from OP who updates on his ongoing project with 'AI-safe: An assessment design tool for the safe use and against the misuse of Gen AI.' The focus of the project is about assessment validity and reliability but not about AI integrity. Information on the applied findings are at the aisafe site. This has a 'thinking tool' for use by vocational education teachers that was user-friendly, efficient and effective. AIsafe is not for or against AI in teaching, it provides a practical solution by focusing on non-exam assessments. Defined AI misuse and AI safe design. What is being assessed, how and why are the basic principles whether AI is used or not. Concepts include context-specific, authentic, collaborative, process driven or generative (as in new information being created). Provided examples of the importance of the concepts.
After lunch, I continue in the Education stream.
I begin by presenting the outcomes of the many projects undertaken at Ara from mid-2023 to the present with 'AI in vocational education: pedagogical support vs academic literacies'. I overview the many projects taken - most drawing on the work of Sharples (2023) whereby AI plays a role to support socially constructivist learning. From the many projects, from foundation through to degree learning and across many disciplines, the main purposes of introducing and integrating AI include modelling industry practice, using AI to support pedagogy and/or to support academic literacies. However, there is a natural tension between pedagogy and academic literacy skill attainment as using AI to 'do the learning' may mean practicing academic writing skills is compromised. Therefore, important to have clear learning outcomes, with regard to using AI to support teaching and learning.
Then Marie-Louise Barry (Ara but now at University of Canterbury) and Gus Walken (Ara) present their work 'reflection on using Gen AI in tertiary education: the case of a project management and marketing course. Summarised the learning activities using AI in project management and marketing. Project management level 6 needed to be updated, especially for currency. Used ChatGPT to develop new resources, templates etc. Learning how to prompt and fact checking required and subject expertise is imperative. Also used AI to create quizzes using XML for Moodle :) Gus described how he used ChatGPT to engage learners in a Level 6 customer engagement marketing. Asked ChatGPT to suggest 4 padlet active learning activities. Selected best, implemented and edited version. Ran this and recorded the class. Ask ChatGPT to identify any student misunderstandings from transcriptions of the recording. Then used ChatGPT to create a quiz for students to revise from analysis of the transcript.
Following is Michelle Simbuland, Jaikaran Narula and Nick Cordery from Eastern Institute of Technology (EIT) who share their work on 'teachers' perceptions on the validity, reliability, and fairness of assessed tutorial as an assessment method: addressing AI-related challenges in designing assessments.' Assessed tutorials used as classrooms unavailable due to closure of the campus from Cyclone Gabrielle. Proposed the used of assessed tutorials as an option to also meet the AI challenge. Assessed tutorials are an oral assessment. Students research the topics covered and share insights and understanding of the tutorial. Peers can also critique answers from others. Students given a list of questions a week before. For each question, there will be a first, second and third responder but the students will not know which questions will be asked and whether they will be first, second or third responder. Session is recorded for moderation. Assessor ask the question and selects the student responders.
Semi structured interviews with kaiako to find out their perspectives of the validity, reliability and fairness of using assessed tutorials and thematically analysed the data. 79% agreed validity can be achieved but questioned if 3 questions sufficient to evidence. Ditto with reliability but requires carefully designed marking rubrics and active cross moderation. 87% agreed that the method is fair, oral assessment increases modality of assessments. Agreement that this method addresses AI related issues. Next step is to collect learner perspective.
The last presentation before afternoon tea is with Vanessa Scholes (Open Polytechnic), Rachel van Gorp (Otago), Jessica Tupou (Victoria University) and Grayson Orr (Otago). They share their work on 'can AI assist us to address cognitive load for neurodivergent online learners? starting our journey. Began by introducing AI, the concepts of cognitive load, neurodivergence and online learning. Tested the hypothesis if brief summaries of online course content could reduce cognitive load for neurodiverse learners. Summarised the key points from the literature review - cognitive load and neurodivergent learners and online learning. AI's potential role to support the challenge and the risks of of using AI. Approaches include learner led - but may be patchy as the learner has to find the tool, work out how to use it and may obtain poor quality outputs. Provider-led will provide tools but quality may still be inconsistent. Educator-led is a balance between but entails more time and work from teachers.
Tested AI tools to see if feeding course content would yield useful summaries. ChatGPT, Claude, Copilot and Brave's Leo AI. Both Leo AI and Copilot had some problems with large amounts of content. ChatGPT seems to produce succinct summaries and Claude also worked but more verbose. Next steps shared - ongoing testing and development, collaboration with learners, teachers and tech to test, more research and evaluation and advocacy and awareness of using AI to support neurodiverse learners.
After afternoon tea, the last group of presentations for the day in the Education stream were presented.
First up, 'advancing pedagogical strategies in Chinese transnational education: lessons from Te Waipounamu with Jeremy Taylor from Otago. Introduced the concept of transnational education - international education delivered in a different country. Over 1200 programmes at degree level but 35 + joint programmes between NZ and China. Study occurred in China (Chengdu and Dalian) with interviews, 8 in each city. Comparison between what the UK does in this space as they have large numbers. Study investigated teaching practices and learner motivations in business management in Chengdu and mechanical engineering in Dalian. Participants were learners, teachers and programme managers. Push/pull factors established for how learners select/or have selected for them, the programmes they enrol in.
Then John Howse from Toi Ohomai sharing an aspect of his PhD with 'exploring vocational education and training's role in just transitions: a practice-based approach to researching VET through a case study in apiculture. Began on the way 'transitions' fit into the vocational educational system. VET is couched as an economical requirement. However, VET's constant change, makes it difficult to keep to the fundamental objectives. Summarised how theory of practice architecture applied to explore VET practices. Applied the theory beekeeping practice. Worked as a bee keeper and completed the studying of bee keeping in a formal institution. Compared his experiences with what was observed at work and what the qualifications required. Found qualifications can not encompass the nuances of practice, especially 'the affective dimension of care' which underpin bee keeping. Summarised plans to continue similar studies across other industries. Stressed the importance of connecting authentic practice to the qualifications.
OP's Rachel McManamara, Amy Benians and Helen Mataiti share their work on 'planning a practice focused inquiry in Universal Design for Learning utilisation: questions we asked ourselves. UDL used mostly in the compulsory sector and there is a need to understand how UDL can be better aligned to VET. Project still in planning stage but as the conversations came about, the questions that arose became important to also analyse. Questions were looked at individually (thematic, linguistic and discourse) and to identify commonalities and differences. What does each know already? How to narrow the analysis tool? Where does UDL sit? What are the strengths in each researcher to see things differently? what is not being seen? A focus on this initial practice, has build relationships and shared values.
Last session for the day is with Neeru Choudhary and Muhammad Arsian from Open Polytechnic with 'AI for higher education - trends, future challenges, and opportunities. Systematic literature review to identify gap to inform future research. AI is not new, perhaps since 1943 but Gen AI in 2023 has popularised the concept. AI in education used in profiling and prediction, intelligent tutor systems, assessment and evaluation and adaptive systems and personalisation. Rationalised the use of systematic review. Inclusion and exclusion criteria important - last 10 years, written in English and peer-reviewed in higher education. Used PRISMA methodology to refine to 246 papers. Current AI technologies include personalised leanring, administrative efficiency and language learning support. Future trends include, high precision education advanced learner tracking, collaborative learning environments and development of sophisticated educational platforms. Challenges are the ethics, data privacy/security and the digital divide. Opportunities to improve educational outcomes, personalised student support, reduced administrative burdens and transformin the educational landscape. Important to better understand student performance and engagement when AI is introduced. Quantitative study planned.
A busy day with many good presentations to unpack.
No comments:
Post a Comment