The design and development of the AI HR manager, powered by Torrino, Torrens University Australia’s (TUA) in-house large language model (LLM), in PSY101 involved a multi-faceted approach, ensuring a robust and engaging learning experience for first-year students in the Bachelor of Psychological Science. In this project, the synergy between cross-functional teams involving subject-matter experts, learning experience designers, GenAI and XR developers, and other stakeholders has been a key determinant of the project’s success. In this subject, our collaborative work incorporated innovative approaches to our assessment strategies and efficiencies in the development of module content, formative learning activities, exciting digital media assets, and an AI-enabled XR environment that facilitates a simulation of real workplace consultation between a student assuming the role of an employee consulting with the Torrino agent acting as an HR manager.
The assessment redesign and the integration of GenAI and XR in PSY101 are grounded in established learning theories and frameworks. Piaget’s (1964) constructivist learning theory informed the approach by positioning learners as active participants who draw on prior knowledge to develop new understanding through interaction with the learning environment. Kolb’s (1984) experiential learning framework guided the simulation component of the assessment, ensuring that real-world scenarios and structured interactions enable our target learners to construct new meaning and insights through concrete experiences, reflection, and active experimentation. In addition, we aligned our innovative approach with TUA’s Learning and Teaching Philosophy and Principles (Torrens University Australia, 2024) to ensure students experience authentic, relevant, and sustainable learning. This initiative supports the imperative to maintain a Psychology curriculum that remains responsive to the evolving digital landscape and the dynamic demands of contemporary workplaces.
In this project, we considered a recent publication on the validity lens for summative assessments, which envisions future workplaces where AI is ubiquitous and ensures learning outcomes are demonstrated sufficiently (Dawson et al., 2024; Elis & Lodge, 2024; Lodge et al., 2023). This perspective aligns with the World Economic Forum’s (2025) Future of Job Report, which predicts increasing demand for GenAI skills in the workforce and recognises the change in the proportion of human-performed tasks due to increasing automation. Next, we also consulted the AI Assessment Scale (AIAS) Framework Level 4, which encourages students to work with GenAI to complete tasks that require human evaluation (Perkins et al., 2024). Finally, our decision to integrate GenAI into a summative assessment was influenced by the recent findings indicating that GenAI detection tools often report false positives (Hardie et al., 2024; Lodge et al., 2023).
The first assessment in PSY101 involves an AI-driven consultation in which students access an XR environment and interact with an avatar representing an HR manager, a Torrino agent named Sam. Students must develop five (5) question prompts aimed at enhancing workplace mental health and wellbeing, addressing areas such as autonomy and role clarity, workload and hours, and professional development. They are then required to critically reflect on their interaction with the AI HR manager, evaluate the accuracy of the AI’s responses, and articulate how these responses connect with, extend, or challenge their understanding of organisational psychology theories and their application to real-world workplace engagement and wellbeing issues.
To scaffold the learning experience of students and prepare them for the first assessment, the learning experience designer, subject-matter expert, and the digital media team collaborated to create comprehensive subject content, including resources, learning activities, and digital media assets such as animated infographics, screencasts, and an interview with an industry expert. This endeavour required the learning experience designer to facilitate subject development workshops and digital media meetings to gather and organise all necessary resources to ensure students are supported in achieving the following learning outcomes:
Explain the key concepts, theories, and factors in organisational psychology in relation to workplace wellbeing.
Investigate the skills required to address wellbeing issues in the workplace and how they apply in contextualised workplace scenarios.
Critically reflect on organisational issues and their impact on worker mental health and wellbeing.
Examine the effectiveness of different strategies employed to facilitate worker and workplace mental health, well-being, and engagement.
To design and develop Sam, we adapted key insights from Tsatsaronis’ (2024) HERDSA presentation and used a customised agent prompt template to contextualise the Torrino agent, enabling it to simulate the role of an HR manager and provide students with an interactive learning experience. This process involved defining the knowledge, expertise, and role of the Torrino agent, tailoring its persona, and structuring predicted interactions and scenarios based on the learning contexts of the target cohort. To guide the behaviour of the Torrino agent and accurately mimic the role of an HR manager, we applied prompt engineering frameworks such as the Goal, Context, Expectations, and Source (GCES) framework (Microsoft, n.d.) and the Context, Objective, Style, Tone, Audience, and Response (CO-STAR) approach (GovTech Data Science & AI Division, 2023). To ensure the AI HR manager communicates professionally and relates effectively to our target learners, we applied Grice’s maxims, which are rules that guide effective communication, including quantity, quality, relation, and manner, as cited in Renkema and Schubert’s (2018) book. If students’ questions fall outside the scope of an HR manager’s role and expertise, the model is programmed to decline politely, prompting the user to rephrase the question so it aligns with the subject matter.
In this project, we successfully addressed copyright challenges associated with scraping PSY101 module resources to create datasets for training the AI HR Manager. We consulted TUA’s Copyright Officer and obtained explicit permission from authors of current, relevant, peer-reviewed journal articles to ensure their authoritative content was accurately represented in the AI. Copyright owners can be assured that the Torrino agent in the Organisational Psychology subject is highly customised, robust, curated, inclusive, and equipped with data protection features. These measures mitigate risks of misuse and misrepresentation of academic content while reducing biases commonly found in publicly available AI tools. Additionally, we strategically limited scraping to resources covered in Modules 1 to 3 to prevent misuse and foster students’ metacognitive skills in regulating reliance on the Torrino agent.
This paper highlights the following primary goals of the development of the AI HR manager with an avatar:
Mitigate the Risks of GenAI: Implement strategies to preserve the integrity of summative assessments and teach students how to ethically and responsibly use AI.
Develop Critical and Analytical Thinking Skills: Equip students to analyse and evaluate AI-generated information critically.
Design Valid and Reliable Summative Assessments: Ensure assessments are robust, valid, and reliable, and prepare students for a future where AI is ubiquitous (Lodge et al., 2023).
Immersive Learning Experiences through XR: Simulate realistic consultations between an employee and an HR manager as a transformative extension of workplace scenarios within educational settings.
Personalised Feedback and Responses through AI: Offer highly contextualised responses, instant feedback, and adaptive learning pathways that allow students to apply organisational psychology theories at their own pace and focus on areas needing improvement.
This paper aligns with the AECT 2025 Convention’s theme “Refocus: The Future of Educational Technology” by showcasing how GenAI can be integrated into a summative assessment to enhance learning and assessment. It underscores the importance of collaboration among stakeholders and explores both the benefits and risks of AI in education.
Through this paper, we aim to inspire educators and institutions to adopt innovative approaches that integrate technology with pedagogy, ultimately preparing students for a future where AI plays a pivotal role in the workplace. As a future research direction, this project provides TUA and other like-minded institutions with an opportunity to contribute to the scholarship of teaching and learning by investigating how AI-XR environments enhance cognition, emotion, behaviour, and learner engagement. According to Ning et al. (2025), the convergence of generative AI and XR to create intelligent, context-aware systems that enhance learning experiences across knowledge domains, along with their ethical considerations and sustainable adoption, represents an emerging area of research worth exploring.
EdTech Archives