Human-AI collaboration is likely to become the norm in the future, meaning that the use of AI tools is likely to become essential for students as soon as they step out of university and into the work field. Moreover, with AI tools becoming increasingly powerful, the measures taken to marginalise AI in assessment may quickly become obsolete. Under these considerations, a possible approach would be to allow the use of AI and even make it an integral part of assessment.

Proponents of allowing the use of AI in assessment argue that students are more likely to approach assessments with integrity when they feel a sense of autonomy, competence, and relatedness. Prohibiting AI may negatively affect them as they may feel less empowered to make decisions about their learning process. Some teachers report that if students are allowed to use AI, they learn through experience to recognise its limitations and to make responsible use of it. We advise doing so only after supporting students in developing some basic AI literacy in the classroom.

The NPULS publication ‘Smarter Education with AI’ (pp.30-32) reports a case of a teacher enabling AI in exams in an eighth-grade high school. Beforehand, they learned about the advantages and disadvantages of different AI systems in class. The teacher reports that the students were reflective when dealing with AI texts and see AI tools more as support than as homework automation.

Other teachers in favour of using AI may opt for making students’ ‘collaboration’ with AI in creating the assessment product an integral part of the assessment. Examples include:

AI use in literature search: Students are allowed to use ChatGPT to find literature for the assignment. As part of the assessment, they must provide their interaction with ChatGPT and critically analyse its responses, highlighting to what extent it contributed to finding the key sources for the assignment against their search using other non-AI tools (e.g., Google Scholar).
Improving an AI product: Students design prompts to have AI draft a product (e.g. policy briefing, draft advice, pitch deck, marketing copy, impact analysis, etc) and improve upon it. They document the process and reasoning: how they crafted and refined the prompts, what choices they had to make to improve the product, etc. The documented process demonstrates learning, is graded, and is more heavily weighted than the product.
AI-assisted data analysis: Students use AI tools to analyse a dataset. They must document how they used the AI tool, the insights it provided, and a critical reflection on its accuracy and limitations compared to their own analysis and interpretation. This documentation will be graded.
AI-assisted coding: Instead of asking students to generate code in the exam, they are presented with AI-generated code and asked to fix it. They must document and motivate their analysis and decision-making process. The documentation is graded.

In these examples, AI serves as a tool within the assessment process. The interaction with the tool can serve as a testing ground for students’ knowledge and skills. For instance, to critically evaluate AI’s suggestions, students must apply their domain knowledge, to improve on AI’s literature search students must use and reflect on their research skills, etc. Integrating AI into the assessment process and guiding students’ reflection can lead not only to more authentic and motivating assessments for students but also to better learning.

🔗 Useful resources

The University of Sydney has various interesting web pages discussing a pro-AI approach in learning and assessment, for example: