Faculty/module/cohort:
Institution-wide perspective, with examples drawn from a range of disciplines including Nursing and English.
AI Tools Used:
Microsoft Copilot (primary), ChatGPT (previously used), ElevenLabs, Blackboard AI Design Assistant, Blackboard AI Conversation tool, and Canva (referenced).
AI Modes Used:
- Text → Text (research, drafting, analysis)
- Voice → Text (dictation and transcription)
- Text → Voice (audio playback)
- Discussion of text → image and text → video as potential assessment formats
What was the challenge?
As a Learning Technologist with responsibilities relating to digital accessibility, Al described several interconnected challenges surrounding the adoption of generative AI within higher education.
Rather than focusing on a single teaching activity, he discussed broader institutional questions concerning how AI can support staff productivity, improve accessibility, and encourage assessment approaches that extend beyond traditional essay-based formats. He also highlighted practical considerations relating to staff development, digital inclusion, and assessment design, noting that institutional processes and assessment practices have evolved more gradually than AI technologies themselves.
Throughout the interview, Al consistently emphasised that decisions about AI should be guided by intended learning outcomes rather than by the capabilities of particular technologies.
What did they do?
Al described a range of established professional practices alongside advisory work supporting academic staff across the university.
He explained that he now routinely uses Microsoft Copilot in place of conventional web searches, describing it as his preferred tool because it operates within the University’s Microsoft tenancy and incorporates restrictions around reproducing copyrighted material.
For staff communications, he described using Copilot to produce initial drafts of blog posts by supplying transcripts of Microsoft Teams conversations with colleagues. These AI-generated drafts are subsequently reviewed, edited and refined before publication.
Al also described an established voice-based workflow that developed after a shoulder injury limited his ability to type. He regularly records spoken notes while walking, transcribes them into text, refines the resulting transcript using Copilot and, on occasion, uses Eleven Labs to convert the edited text back into speech for review.
He further reported using Copilot to assist with the analysis of anonymised student engagement data, particularly in situations where statistical software would otherwise be required.
Beyond his own practice, Al described supporting staff in the use of Blackboard’s AI tools. These included the AI Design Assistant, which can assist tutors in producing quiz questions and discussion prompts, and the AI Conversation tool, which enables educators to create conversational AI personas for role-play activities. Examples discussed included simulated workplace scenarios such as organisational mergers and interactions designed to support learning within healthcare education.
Multimodal Use
The practices described combine multiple AI modalities to support different aspects of teaching, learning and professional work.
Within Al’s personal workflow, spoken dictation is converted into text through transcription, refined using text-based AI and, where appropriate, converted back into audio for review. This creates an iterative workflow that combines voice and text throughout the drafting process.
At an institutional level, Al described Blackboard’s AI Conversation tool as supporting conversational learning through sustained dialogue with AI-generated personas rather than relying solely on conventional written assessments.
He also discussed the potential for multimodal assessment formats—including infographics, presentations and video—to provide alternative methods of demonstrating learning, provided that assessment continues to evaluate the intended learning outcomes rather than students’ proficiency with particular software tools.
What did this look like in practice?
Al described several examples illustrating how AI has been incorporated into educational practice.
Within Nursing education, he referred to assessments requiring students to communicate public health messages through posters or infographics. He suggested that AI-supported design tools such as Canva or Copilot may enable students with limited graphic design experience to produce professional-quality materials while still demonstrating the intended learning outcome of effective health communication.
He also described Blackboard’s AI Conversation tool being used to simulate professional scenarios, such as organisational mergers, where students adopt workplace roles and develop their understanding through ongoing dialogue with an AI persona over an extended period.
His own professional workflow similarly illustrates multimodal AI use, combining voice recording, automated transcription, AI-assisted text refinement and audio playback into a single drafting process.
What was the impact?
Al reported that AI has had a positive impact on his own professional practice by improving efficiency across routine tasks. He explained that AI enables him to complete work more quickly and maintain focus on activities that he might otherwise postpone. He related this observation to his own experience of neurodivergence, describing AI as a useful cognitive support rather than a replacement for his work.
From an institutional perspective, Al identified several potential benefits associated with more flexible assessment design. He suggested that offering alternative assessment formats could improve accessibility for some students, particularly those who experience significant anxiety around assessed presentations or similar forms of performance.
However, Al also noted that he had seen relatively limited evidence of widespread adoption of genuinely multimodal assessment across the university. In his experience, essays continue to dominate assessment practice, while technology is often incorporated primarily to strengthen existing assessment processes rather than fundamentally changing how learning is demonstrated.
Ethical or practical considerations
Al identified several practical and ethical issues associated with the expanding use of generative AI.
Digital inclusion
He emphasised that students do not necessarily begin university with equivalent access to technology or comparable levels of digital confidence. Examples included differences in access to personal computers and variation in fundamental digital skills, which he suggested should be considered when introducing technology-enhanced assessments.
Staff development
Al also observed variation in AI knowledge and training among academic staff, noting that differences in confidence and experience may influence how AI is adopted across different programmes.
Data privacy
He expressed concern that students may interact with AI systems as though they were conventional search engines without fully considering the implications of sharing personal or sensitive information.
Reliability of AI detection
Al explained that the University of Northampton does not currently rely on Turnitin’s AI-detection functionality because internal evaluation indicated that its results were insufficiently reliable for academic misconduct investigations. He also noted that advances in generative AI have made it increasingly difficult to distinguish AI-generated writing from an individual’s established writing style.
Fairness in academic integrity processes
He described how colleagues within the English department had moved away from treating individual words or phrases as indicators of AI use, concluding that this approach was neither sufficiently reliable nor equitable.
Alignment between assessment and learning outcomes
Finally, Al questioned assessment approaches that require students to demonstrate authenticity through unfamiliar technologies if competence with those technologies is not itself an intended learning outcome.
Reflections and Advice
What worked well?
Al described AI as being most valuable when used as a cognitive support rather than a replacement for thinking. He highlighted its usefulness in refining ideas, supporting drafting processes and enabling students to demonstrate knowledge without being constrained by unrelated technical skills such as graphic design.
What would you refine?
Looking ahead, Al suggested that assessment design could place greater emphasis on documenting students’ learning processes, including planning, drafting and iterative development, rather than evaluating only final outputs. He illustrated this through an analogy with portfolio-based assessment in creative disciplines, where the development process forms an important part of assessment.
Advice for colleagues
Al recommended beginning with intended learning outcomes before considering the role of AI. He advised colleagues to avoid adopting technologies primarily because they appear innovative, arguing instead that AI should be introduced only where it demonstrably supports learning, accessibility or assessment.
Further reflections
Throughout the interview, Al suggested that many of the current challenges surrounding AI are institutional rather than technological. In his view, questions concerning assessment design, staff development and digital inclusion will be as important as technological capability in determining how AI is integrated into higher education.
Quick Start Guide for Colleagues
Estimated time required:
Low for individual AI-supported workflows such as drafting and transcription.
Moderate to High for designing AI-supported role-play activities or alternative assessment formats using institutional platforms.
Digital skill level required:
- Students: Low to Moderate
- Staff: Moderate
Common pitfalls to avoid:
- Assuming all students have comparable access to technology or equivalent levels of digital confidence.
- Introducing technology that is not directly aligned with the intended learning outcomes.
- Relying solely on AI-detection software when making academic integrity decisions.
- Selecting AI tools because of novelty rather than pedagogical value.
Best suited for:
Institution-wide digital education initiatives, health and communication-based disciplines, employability-focused teaching, professional role-play activities, and programmes exploring alternative assessment approaches beyond traditional essays.
