Generative AI and assessment at AUT

AUT takes a whole-of-institution approach to the systematic integration of generative AI (Gen AI) into assessment design. This work is grounded in the assessment principles, policy and procedures that provide a foundation on which to build a sustainable approach to the integration of Gen AI into assessment and feedback practices.

The impact of Gen AI on tertiary education has been significant. While Gen AI has created new possibilities for learning and teaching, it has also highlighted shortcomings in current assessment design in tertiary education. The key challenge for universities is how to assure that the learning that has taken place reflects the expected outcomes of a student graduating in a particular discipline.

Two AUT assessment principles in particular reflect the presence Gen AI:

1.
I am prepared for life / Whakaritengā mo te Oranga

Assessment will equip students for learning throughout life.

By engaging with authentic assessment tasks that have value beyond their programme, students are better prepared for continuous learning in the world of work, with their whānau and communities. When encountering new challenges, such as emerging technologies, they have a sense of agency and act with integrity.

2.
I am clear about the purpose of assessment and my expected performance / Te Kaupapa me ngā Tūmanako

The purpose of a task, its relationship to the learning outcomes and how it will be assessed are explicit.

Students will be able to recognise the intention of an assessment and its connection to the learning outcomes and teaching activities. Assessment criteria identify the required performance standard, and opportunities are provided to discuss the requirements of a task. Students will be clear on how technologies can be used to complete each assessment.

Assessment Redesign for Engaging with AI (AREA)

AUT’s approach to designing assessments with Gen AI has drawn heavily on the work of the University of Sydney and the pioneering work of Danny Liu and colleagues.

AUT has adopted Sydney’s two-lane model to the use of Gen AI in assessment and has reframed this so that assessment sits within one of two channels:

Channel 1 - Secured, controlled assessment tasks where it may be decided that generative AI tools cannot be used
Channel 2 - All other assessments acknowledge the ubiquity of AI, and its use is permitted within academic integrity guidelines and clearly communicated to students. The extent of the use of AI in an assessment will vary, depending on the assessment design

For assessments to sit authentically in Channel 2, it is likely that a redesign of assessment tasks will be required. LTED has created guidance on Assessment Redesign for Engaging with AI (AREA).

Why we need to redesign assessments for Gen AI

While some forms of assessment may be better able to withstand the threat posed by Gen AI, many forms of assessment tasks are at risk. There is a need to review assessments and, where they are at risk, redesign them.

What it means to redesign assessments with Gen AI

We need to engage openly, critically and ethically with Gen AI as a part of the student learning experience, including thoughtfully designing assessments where AI can be used in different assessment tasks.

LTED | Office of Learning, Teaching and Educational Design

LTED is a central AUT service with a team of experienced and dedicated learning designers, educational video specialists and learning technologists.

Assessment design

AUT aims to equip students with knowledge and skills to take them beyond their time at university. A key driver in achieving this objective are AUT’s assessment principles, policy and procedures.

Resources for AUT staff

AUT staff can access more information and guidance on generative AI in assessment on TUIA, including what support is available to help staff redesign assessments in line with AUT’s approach.

GEN AI HUB ON TUIA