Are we zombies? Validity is the word for assessment practitioners

10 November 2025

In this post we hear PhD student Lhea Reinhold’s thoughts on our latest webinar ‘Secure assessment tasks in a time of GenAI’. This webinar was the third in the ‘New Directions in AI Research and Practice’ series. Lhea is currently visiting CRADLE from Friedrich-Alexander University Erlangen-Nuremberg, Germany, and is working on her dissertation ‘The Role of the Teacher in AI-supported Assessment Formats in Quaternary Education’. 

Webinar #3: Secure assessment tasks in a time of GenAI

Last month CRADLE hosted the final webinar in our New Directions in AI Research and Practice series. Facilitated by Dr Thomas Corbin the webinar featured discussions around securing assessment with Professor Phillip Dawson (CRADLE), Professor Cath Ellis (Western Sydney University), Professor Danny Liu (The University of Sydney), and Kane Murdoch (Macquarie University).

In this post I talk about the panellists’ perspectives on securing assessment tasks in a time of GenAI. It is a highly important topic to discuss, in order not to become “zombies” that are walking around headless and bringing more harm than good to assessment and learning processes. 

1. A call for assessing

To capture the spirit of the webinar, it is important to address the “discursive changes” (Corbin et al. 2025a) within assessment processes by using the right vocabulary first. While words shape normality, it is of major importance to shift our attention from the noun “assessment” to the verb “to assess”. Assessment practises should be enacted by the learner, as well as inspire and assure an active learning process.

2. A call for trade-offs

In today’s world, assessment tasks are under immense pressure: GenAI is able to support learners in their assessments, and therefore the following question – not new, but more urgent than ever before – arises:

What makes an assessment task valid, secured, and authentic?

This complex question cannot be answered within one webinar, blog entry or through a one-size-fits-all solution. However, to address this “wicked problem” (Corbin et al. 2025b), the panellists shared the following guidance:  

• Ask yourself what the main purpose of the assessment is. 

• Then ask yourself what assessment task could support the indented purpose of the assessment. 

By weighing aspects such as validity, security, and authenticity within assessment designs, practitioners get a chance to handle the current situation. 

3. A call for practitioners

In case a practitioner aims to design a more valid assessment task and therefore to be confident that the task sufficiently measures the intended learning outcome, Professor Dawson offered the following advice: validity is not an absolute concept for psychologist, nor can it be proven. Validity involves finding individuals points of evidence to support judgments about student capability and, therefore, it is a concept enacted by practitioners. 

How does this influence my research?

In 2025, I was deeply interested in personalizing assessment, to what extent such personalized assessments can be considered valid, and how far they reshape the relationship between standardization and personalization. For my empirical study, I prepared several single points of evidence, as Professor Dawson emphasized. While listening carefully to the experts, I realized that I had not been balancing the different aspects related to assessment; rather, I had focused solely on validity. As a result, I intend to revisit my research and approach it with greater nuance. 

In addition, I am very grateful for all the new articles shared by the experts and participants – I had a very focussed reading session after the webinar. 

About Lhea Reinhold

Lhea is interested in adult education, assessment formats and generative artificial intelligence. With the aim of leveraging scientific insights, Lhea always ensures a practice-oriented research approach and strives to implement her findings within the continuing education sector.

• LinkedIn

Catch up on the whole series

References

Corbin, T., Dawson, P., & Liu, D. (2025a). Talk is cheap: why structural assessment changes are needed for a time of GenAI. Assessment & Evaluation in Higher Education. Advance Online Publication. https://doi.org/10.1080/02602938.2025.2503964

Corbin, T., Bearman, M., Boud, D., & Dawson, P. (2025b). The wicked problem of AI and assessment. Assessment & Evaluation in Higher Education. Advance Online Publication. https://doi.org/10.1080/02602938.2025.2553340

CRADLE has been busy researching GenAI – find our latest publications on our blog.

• Latest GenAI publications

CRADLE Seminar Series

More than the individual: transforming feedback cultures

Tuesday 11 November 2025 at 2pm

• Register now!

• YouTube
• Bluesky
• Email CRADLE
• CRADLE Website