Assessment and AI

AI in Education: A Realist’s View on Assessment and Assurance

In the rapidly evolving world of technology, Artificial Intelligence (AI) has taken centre stage, and education is no exception. With the integration of AI systems – like chatbots into educational platforms – new opportunities and challenges arise, especially in the context of assessment.

The Tempting Texts of AI

One of the many capabilities of AI is the production of highly engaging and responsive text. The allure is undeniable, but these systems are not without limitations and biases. Soon, AI chatbots will be integrated into common tools like Microsoft Office, which may address intellectual property issues.

The Challenges of AI in Assessment

AI’s integration into educational assessment is complex. Its use may challenge the validity of student evaluations, raising questions regarding how to ensure that students meet appropriate learning outcomes. Additionally, technology creates new aids that students and educators alike will adopt, creating both new affordances and hurdles.

A Closer Look at Assessment

Assessment in education serves three key purposes:

  1. Assurance: To guarantee that learning outcomes have been met, conforming to the standards set within universities.
  2. Enablement: To provide students with information to aid their learning and meet the desired outcomes.
  3. Capacity Building: To help students address their own learning independently, beyond the educational institution.
Assurance in Assessment

Assurance aims to judge whether students demonstrate attainment of learning outcomes to a given standard. It requires transparent standards and does not allow for judging students against each other. Here are some approaches to tackling the challenges of AI:

  • Nested Assessment: Sequential assessment tasks allow for monitoring continuity in a student’s work, potentially highlighting inconsistencies.
  • Task Design: Designing assignments that cannot be easily answered by AI, focusing on highly contextualized, personalized, and recent tasks.
  • Programmatic Assessment: Focusing on key learning outcomes in a program rather than individual courses.
Enabling Learning Through Assessment

Feedback has undergone a transformation in understanding; it’s now seen as a process, not merely an input. It must be designed to engage the student actively, and it should be separated from grading. Assessment for learning helps students engage with the content, and if done correctly, enhances their success in assurance tasks.

Building Students’ Judgement Capacity

Assessment must also develop students’ ability to judge the quality of their own work. This involves focusing on quality and providing opportunities for self-assessment and peer-assessment. It enables them to use AI tools effectively by developing their skills in prompting and judging responses.

Concluding Thoughts

The integration of AI into education is neither wholly positive nor negative, but rather a nuanced issue demanding careful consideration and innovation. A regression to traditional exams is not the answer; it may even undermine standards. Instead, a major redesign of assessment methods is required, focusing on unique, specific, current, and contextualized tasks.

The future of education with AI requires more assurance for learning and fewer but more complex tasks relating to threshold standards. It’s about securing a few key assessments and assuring the program as a whole, not just individual modules.

The interplay of AI and education is still unfolding, and as technology continues to evolve, so must our approaches to teaching and assessing. The goal is to provide education that is both robust and adaptable to the ever-changing landscape of technology.

Finally, it’s important to recognise that all educators and institutions are likely to have their own unique opinions on the integration of AI. As a result, this integration must be handled and regulated at a level higher than that of the individual educator. Taking the steps discussed above, however, is a good starting point to ensure that the correct approach is taken in this world of AI.

This post summarises a lecture by David Boud on Positioning assessment differently in a world of Gen AI, hosted by Digitally Enhanced Education Webinars at the University of Kent. The lecture is available on YouTube.

#AIinEducation, #genAI, #AssessmentChallenges, #StudentLearning, #AssuranceinLearning, #FeedbackProcess, #IntellectualProperty, #AcademicIntegrity, #ProgrammaticAssessment, #TraditionalExams, #TechnologyinAssessment

Amir-Homayoun Javadi

Amir-Homayoun Javadi, PhD

Founder and CEO at 0&1 LTD