Reimagining assessment in the age of generative AI—The strategic role of Cadmus

Overview

Generative AI is reshaping how we engage with knowledge, expression, and cognition. For higher education, this transformation presents both significant risks and far-reaching opportunities. At Cadmus, we believe the future of assessment is the future of learning—and that generative AI has the potential to fundamentally change how assessment supports intellectual development and educational integrity.

Cadmus is recognised as a sector leader in written assessment. Our expertise in this space has been built through deep partnerships with educators, a commitment to academic integrity, and platform design that foregrounds learning. As generative AI becomes more integrated into academic contexts, it is precisely through writing—where students must structure, defend, and reflect on their thinking—that its influence will be most profound.

While Cadmus already supports a wide range of assessment formats—including multi-format assessments including quizzes, short- and long-answer tasks, time-bound exams, and collaborative group work—our future focus is on expanding into new and diverse assessment modalities. These include oral assessments and vivas, programmatic assessments, multimodal submissions, and authentic, real-world tasks. This evolution is grounded in the same pedagogical principles that have defined our success to date: academic integrity, learning outcomes, and student-centred design.

Written assessment: The critical intersection of learning and AI

Assessment has always been more than measurement—it is where much of the deepest student learning takes place. In the age of generative AI, this becomes even more pronounced. Because these technologies operate through language, writing becomes the primary mode of interaction between students and AI tools.

This reinforces rather than diminishes the value of written assessment. Through writing, students engage in complex cognitive work: planning, reasoning, articulating arguments, and evaluating information. In this context, Cadmus serves not only as a platform for delivering assessments but as a strategic site for understanding and guiding how students learn in AI-enabled environments.

Empowering educators through AI-augmented workflows

The transformative potential of generative AI extends beyond students to the academic workforce itself. By streamlining repetitive or administrative tasks, AI can enable educators to reclaim time and energy for teaching, mentoring, and curriculum innovation.

Cadmus will continue to develop tools that reduce burdens across the assessment lifecycle, including:

• marking and feedback generation
• moderation and calibration processes
• assessment management workflows
• academic integrity triage and investigations
• student engagement and success optimisation.

Through these efficiencies, we aim to elevate the educator’s role, not automate it—allowing academics to focus on advancing learning and guiding students through increasingly complex educational terrain.

Supporting, not replacing, student cognition

For students, generative AI presents new opportunities to support—not supplant—their academic development. When integrated thoughtfully, AI can enhance the research, preparation, and reflective dimensions of assessment by assisting with:

• compiling relevant academic literature for review and exploration
• structuring learning frameworks that align with disciplinary conventions and best-practice academic processes
• providing academic skills support, including referencing guidance, critical reading strategies, and time management
• offering grammar, spelling, and communication feedback, across both written and spoken tasks
• delivering formative suggestions or prompts that encourage deeper engagement and stimulate intrinsic motivation to persist with complex tasks.

Crucially, these functions must be designed to enhance student agency and cognitive engagement, not replace them. AI should serve as a scaffold, helping learners develop confidence, curiosity, and autonomy across multiple modes of expression.

A recent article from the MIT Schwarzman College of Computing cautions against the overuse of generative AI tools such as ChatGPT, warning that such reliance may lead to cognitive atrophy—where students forgo the effortful processes of thinking, planning, and reflection that underpin deep learning.

At Cadmus, our commitment is to design AI-enabled environments that reinforce metacognitive and intellectual development—whether students are writing, speaking, or creating. Our platform will continue to foreground intentional, transparent, and academically rigorous use of AI.

Shifting the integrity conversation: From detection to motivation

In a learning landscape transformed by generative AI, it is no longer sufficient—or ethical—to rely on monitoring and detection as the primary mechanisms for upholding academic integrity. While AI detection tools may be technically scalable, they are not pedagogically or ethically defensible. Their high rates of false positives and false negatives—coupled with a lack of transparency and due process—create unacceptable risks for students and institutions alike.

More critically, these tools present a deeper risk: they create the illusion of a solution. AI detectors offer a sense of institutional control and reassurance, but they do not meaningfully work—either in detecting unauthorised use of generative tools or in supporting the educational mission. This false confidence can distract from more constructive and lasting approaches to academic integrity grounded in trust, skill development, and student motivation.

Similarly, the relevance of traditional similarity detection is diminishing rapidly in a world where AI-generated content is often entirely original in structure but lacks academic authenticity or depth.

As we move forward, we will double down on this approach. The true solution lies not in trying to outpace technological misuse, but in cultivating intrinsic motivation—within both students and educators—to engage meaningfully with learning and uphold shared academic values.

Our roadmap will increasingly focus on:

• designing assessment environments that foster ownership, curiosity, and academic pride
• supporting educators in reframing their roles as mentors and facilitators of lifelong learning
• reinforcing the identity of higher education institutions as sources of knowledge, research, and best practice, not simply credential providers
• encouraging students to see their education as something to invest in intellectually and ethically, not just complete.

In doing so, Cadmus aims to help institutions navigate the complexity of the AI era while protecting the foundational mission of higher education: to empower generations of learners and educators to pursue truth, knowledge, and meaningful contribution.