From Classroom Design to Newsroom Practice: Assessment Intervention Designing GenAI

Synthesis: This paper presents a practice-based curriculum intervention that embeds generative AI (GenAI) into an undergraduate journalism Assessment item, framed by an ethical model integrating Colin Beard's experiential learning design with Bloom's revised taxonomy. GenAI promises efficiencies and new creative possibilities for students yet threatens academic integrity, authorship, and the core values of the profession. The triadic values–processes–competencies framework scaffolds students from human-only reporting to AI-integrated work and critical reflection, converting their intuitive concerns into explicit competencies in ethical decision-making, prompt engineering and verification. Qualitative data from a second-year journalism unit at an Australian global campus in Sarawak, Malaysia show students already use GenAI strategically but hold deep concerns about accuracy, originality and disclosure.

Key Findings

• Ethical framework as a scaffold for GenAI integration: A triadic values–processes–competencies model — foregrounding the journalistic values of disclosure, independence, accuracy and integrity — provides an ethical grounding that moves students through Bloom's cognitive anchors (applying → analyzing/evaluating → creating) within Beard's experiential learning cycle.
• Purpose-built context enables meaningful use: Students framed AI as "best support system", "reflect better", "useful for ideas and structure" and "better accuracy," positioning themselves as active participants in the journalism assessment process; all nine participants agreed AI tools were useful in completing their assessment.
• Confidence vs. depth of ethical understanding: Eight of nine students (89%) reported confidence in making ethical decisions, yet survey items on authorship and ethics suggested superficial understanding — most disagreed that GenAI tools carry biases that could undermine news accuracy, reflecting an assumption that technology is unlikely to be biased.
• Improved prompt literacy over ethical depth: The intervention primarily influenced the depth rather than the level of ethical understanding, but it enhanced engagement with prompt literacy — students reported structured prompting strategies including role-playing, verb-first instructions and specifying tone, audience or length.
• Intervention drives higher-order reflection: Experiential learning activities engaged students cognitively and affectively, moving them toward evaluative and reflective judgment; they expressed a stronger sense of responsibility for accuracy and originality and voiced possibilities of becoming curators/editors who supervise algorithmic content rather than producing it from scratch.
• Tension between traditional values and technological utility: Students self-reported concern about reduced professionalism among journalists who use GenAI, reflecting a transformative phase of experiential learning in which they confront previously held assumptions.

Study Design & Method

The study adopts a design-based approach to develop a practice-based Assessment intervention for the second-year core unit Interactive Storytelling within the Bachelor of Media and Communication at an Australian global campus in Sarawak, Malaysia. Nine of 13 enrolled journalism students participated over one academic semester, selected through purposive sampling. The intervention scaffolds movement from human-only reporting to AI-integrated reflection using Beard's experiential learning design, anchored by Anderson & Krathwohl's revised Bloom's taxonomy.

The framework integrates Beard's experiential learning sequence (concrete experience, reflective observation, abstract conceptualization, active experimentation) with Bloom's levels used as "anchors" rather than a strict ladder. Pre- and post-intervention surveys (Qualtrics, ~10 minutes, anonymous and voluntary) were distributed before and after journalism assessment deadlines. Students completed two assessment tasks and four classroom activities (contrasting research on student GenAI use, debating authorship, discussing bias in GenAI image generation, and reviewing AI-generated work presented as human-authored). Responses were coded thematically using in vivo coding. Trustworthiness followed Lincoln and Guba's framework (credibility, transferability, dependability, confirmability), supported by peer assessment, triangulation across two campuses and an audit trail. The framework aligns with Australian quality expectations (TEQSA Higher Education Standards Framework) and institutional practice (Swinburne's Educational Quality Excellence Framework).

What this means for practice

• Instructors. Replace detection-based responses with a designed assessment that embeds GenAI explicitly, and add classroom activities — debating authorship, discussing bias in AI image generation, reviewing AI-written work presented as human-authored — to build the ethical reasoning behind disclosure.
• Instructors. Turn values into gradeable terms: write rubric criteria that express disclosure, accuracy, and independence as discrete acts such as statements addressing conflict of interest or omission of sources.
• Instructors. Prepare students for human-in-the-loop newsroom roles by having them specify source requirements, flag factual uncertainty, and maintain authorship clarity.
• Designers. Reuse the values–processes–competencies triad across disciplines, since the four values and the experiential sequence are deliberately high-level, and adjust unit outcomes, local journalism codes, and tool choice per context.
• Administrators. Embed GenAI assessment design in institutional quality and integrity frameworks rather than treating ethical AI use as a discrete technical skill.

Limitations

• Nine of 13 enrolled students in a single second-year journalism unit over one academic semester at one campus (Australian global campus, Sarawak, Malaysia), so the design-based intervention yields a formative evaluation rather than a generalizable result.
• Data came mainly from reflection, discussion, and pre/post surveys, so students' ethical competencies are self-reported rather than independently measured.
• Eight of nine students (89%) reported confidence in making ethical decisions while survey items on authorship and bias suggested superficial understanding, indicating self-report may overstate ethical competence.
• The authors note limited GenAI integration in Malaysian journalism education and industry, leaving the framework's adaptation to other cultural and regulatory contexts untested.

Citation

Ngu, I.-Y., & Weller, D. (2026). From Classroom Design to Newsroom Practice: Assessment Intervention Designing GenAI . Journal of University Teaching and Learning Practice, 23(5).