Page 19 - Center for Faculty Development and Learning

Assessment and AI by Dr. Gary Williams,

Associate Professor of Digital Transformation, College of

Business Administration

This article explores strategies for integrating Generative Artifi-
cial Intelligence (GenAI) tools into the design and completion of

summative assessments conducted in unobserved environ-
ments. These assessments, typically completed off-campus,
contribute to a student’s course grade.
It is now evident that the majority of students incorporate
GenAI tools (e.g., ChatGPT, Copilot, Claude) into their learning
processes. Encouraging the use of these tools aligns with con-
temporary learning practices. However, this widespread adop-
tion of GenAI raises critical concerns about a fundamental re-
sponsibility of higher education institutions: to certify the knowledge, skills, and competencies of
graduates through the conferral of qualifications.
A growing body of evidence indicates that reliably detecting the use of GenAI in assessments—
either manually by instructors or through detection software—is increasingly unfeasible. As Mol-
lick (2024) observes, the “‘Detection Illusion’ leads educators to rely on outdated assessment
methods, believing they can easily identify AI-generated work when, in reality, the technology
has far surpassed our ability to consistently detect it.” Similarly, Perkins et al. (2024) argue that
“the varying performances of GenAI tools and detectors indicate they cannot currently be recom-
mended for determining academic integrity violations due to accuracy limitations and the poten-
tial for false accusations, which undermine inclusive and fair assessment practices.”
Given these challenges, instructors face two primary options for managing summative assess-
ments:
Option 1: Ensure that students cannot use GenAI by requiring them to complete assess-
ments in observed, on-campus environments with no access to electronic devices or only access
to a locked-down web browser environment.
Option 2: Design assessments that incorporate the facilitated use of GenAI tools.
While observed assessments are critical for certifying a student’s independent knowledge,
skills, and competencies, it is equally important to include non-observed assessments, such as
assignments completed off-campus. These assessments reflect real-world applications of
knowledge and the integration of tools like GenAI into learning processes.
Table 1 outlines strategies for developing and implementing "Option 2" assessments. These
strategies acknowledge the ubiquity of GenAI tools and their growing role in the learning practices
of both students and instructors. By embracing this reality, we can design assessments that pro-
mote academic integrity while leveraging the potential of GenAI to enhance learning outcomes.

17