The opportunities and risks of AI offer a valuable opportunity to rethink assessment design practices. There are different approaches to responding to AI depending on the needs of specific programs and courses. Here we look at program- and course-level considerations, as well as expectations on students.
With increasing capability and availability of generative AI, it is almost impossible to design ‘AI proof’ assessments. We want to prepare our graduates to work in a world where AI may be ubiquitous, and they need to apply their professional- and discipline-specific knowledge to judge AI outputs.
With this new AI reality, there is value in incorporating generative AI into learning. We have an opportunity to develop graduates with AI proficiency and digital literacies, and to further reduce barriers to learning through AI enhanced Assistive Technology (AT).
At the same time, assuring learning outcomes and minimising the threat of AI to assessment integrity is a pressing concern.
Assessment security and academic integrity
In the context of higher education assessment design, assessment security and academic integrity are frequent terms that appear. Assessment security is defined by Dawson (2021) as specific measures taken to harden assessment against attempts to cheat, including approaches to detect and evidence attempts to cheat, and measures to make cheating more difficult. Dawson (2021) contrasts this to academic integrity, which focuses on educating students about the importance of integrity and encouraging them not to cheat.
Program and course level considerations
Program level
At a program level, to ensure appropriate credentialing and tracking of student progress across the overall award, it is important to have a systemic/programmatic approach to assessment (Lodge et al., 2023). This means aligning assessments with overall disciplinary learning outcomes and identifying key assessment moments at a program level that require assessment security. The TEQSA document on Assessment reform for the age of artificial intelligence provides further guidance and examples of this.
Course level
At a course level, assessment tasks can also vary in their levels of permitted AI use. Some assessments require students to achieve learning outcomes without any use of AI, and there are some assessments where AI is less or not applicable. There may also be some assessments that provide opportunities for students to use AI as an integral component of the assessment.
Speaking with students about AI and assessments
Educators are knowledgeable experts in their respective fields and professions. They can play an integral role in guiding students towards what is an ethical, responsible, and valuable use of AI in their discipline. This may involve:
- communicating clear expectations of what is permitted or unpermitted
- designing assessments of varied permitted AI use
- having in-class discussions of AI with relevance to the discipline.
Students also have a key role in taking ownership and accountability of their work and ideas and ensuring that they uphold values of academic integrity. This means that even when permitted to use AI for assessments, they must acknowledge and document their use of AI, check the validity and reliability of AI generated output, and be responsible for any errors in the generated material (AAIN Generative AI Working Group, 2023).
How to plan assessments in response to AI
To best respond to the opportunities and challenges brought by AI, educators can use the following steps as part of their assessment planning process.
Please refer to the ‘Assessment planning Steps’ pages in this collection for further guidance.
This page is a living document that may be updated in the future as the AI landscape continues to evolve. The content in this resource is intended as guidance. Links to relevant ANU policies and procedures are provided where relevant.