Invited Panel: Defending Assessment Security in the Age of Generative AI

Thursday, April 3rd (2:30-3:30 pm)

As artificial intelligence (AI) tools become more sophisticated and accessible, concerns about academic integrity in assessment continue to grow. There is no one-size-fits-all approach to develop strategies to mitigate these concerns. This panel will explore the evolving landscape of AI and cheating, bringing together perspectives from a testing industry expert, a higher education administrator, and a faculty expert on classroom-level challenges. Panelists will examine the impact of AI on assessments, institutional policies and responses, and practical strategies for maintaining academic integrity in the classroom.

The industry expert will address how AI challenges traditional assessment methods and how testing organizations are adapting to ensure the security of their standardized large-scale exams. The higher education administrator will discuss institutional strategies, including policy development, faculty support, and the role of technological solutions in mitigating AI-related cheating. Finally, a faculty expert will provide insight into classroom-level responses, exploring how AI is changing student behaviors and what instructors can do to foster ethical learning.

A discussant will moderate the conversation, ensuring a dynamic exchange of ideas among panelists and engaging the audience through structured Q&A. This session aims to provide attendees with actionable insights to uphold academic integrity in the AI era.

Discussant/Chair: Ahmet Dursun

Panelist 1: Reza Neiriz, MetaMetrics

AI vs. Exams: Securing the Future of Standardized Testing

Abstract: As artificial intelligence rapidly transforms education and assessment, traditional testing methods face unprecedented challenges. In this short talk, I will explore emerging threats posed by AI to exam security, including automated test-taking and answer synthesis. I will also discuss how the industry is responding through enhanced security protocols, AI-driven proctoring, AI-generated item banks, examiner behavior analysis, and innovative assessment designs that prioritize critical thinking over rote memorization. By rethinking how we evaluate knowledge, we can uphold the validity of standardized testing in the AI age, and more importantly, align with the paradigm shift that AI is ushering in this new era.

Bio: Reza Neiriz holds a PhD in applied linguistics and technology with a focus on computer-mediated language assessment. He is a Machine Learning Engineer at MetaMetrics Inc. researching and developing AI-driven assessment solutions. His research focuses on automated assessment of performance tests, especially constructs like interactional competence in tests of oral communication.

Panelist 2:  Brent A. Green , BYU–Hawaii

AI, Cheating, and Validity—An Institutional Perspective

Abstract: As artificial intelligence (AI) reshapes education, concerns about academic integrity have intensified. While cheating remains a critical issue, it is ultimately a subset of a broader concern that Dawson et al. (2024) refer to as assessment validity—the degree to which assessment scores are interpreted and used accurately to reflect student learning. I will argue that institutional responses should focus less on policing AI-related cheating and more on ensuring that AI-era assessment decisions remain valid and meaningful.

Given this shift in focus, I will also discuss ongoing institutional efforts to address these challenges. We are currently exploring strategies to mitigate these concerns, including faculty training in assessment design, guidelines for students on ethical AI use, and policy frameworks that balance security with pedagogical soundness. Rather than reacting solely to misconduct, our goal is to investigate institutional approaches that promote academic integrity while adapting to the evolving role of AI in education.

By reframing the conversation around validity rather than just cheating, institutions can shift from reactive enforcement to a proactive, evidence-based approach that strengthens learning outcomes and academic integrity in the AI era.

Bio: Brent A. Green is the Associate Academic Vice President for Accreditation, Assessment, and Curriculum at Brigham Young University–Hawaii, where he also teaches courses in language testing and English syntax. He holds a Ph.D. in Applied Linguistics and Language Assessment from UCLA and has dedicated over thirty years to teaching, testing, mentoring, and administration in higher education. His work focuses on language assessment, academic accreditation, and educational policy, contributing to the advancement of both assessment practices and language pedagogy. 

Panelist 3:  Nicholas Swinehart, University of Chicago

Classroom Assessment and Generative AI: Defining and Encouraging “Appropriate Use”

Abstract: As AI tools continue their spread into daily life and education, clear and effective guidance on the appropriate use of AI by teachers and students often lags behind. This has the potential to negatively impact several aspects of instruction at the classroom level, including assessment. This presentation discusses several necessary conditions for the effective and appropriate use of generative AI in classroom assessment: adequate AI literacy and assessment literacy, co-construction of policies on AI use, and rethinking assessment constructs in light of the current realities of generative AI.

Bio:  Nicholas Swinehart is the Managing Director of Instructional Technology at the University of Chicago Language Center, where he supports instructors of over fifty languages with technology use and professional development. He is co-editor of the February, 2024 CALICO Journal special issue on social media for language learning and co-author of Teaching Languages in Blended Synchronous Classrooms: A Practical Guide (2020).