Behind the Screens is where we swap ideas, test out new tools, and see what colleagues are experimenting with. This time the spotlight was on AI: could it make course design faster and more collaborative, and how might it reshape assessment? We had a hands-on demo of Coursensu and considered a thought-provoking model for portfolio-based assessment.
Demo 1: Larisa and Tim – designing courses with Coursensu
Larisa Grice, Senior Learning Designer, walked us through their pilot of Coursensu, a learning design platform built around Diana Laurillard’s six learning types. The tool is designed to support collaborative course planning while integrating AI support, visual design, and convenient export options. It’s also an opportunity to look at a light-touch approach to Curriculum Management.
Key features
- Visual course mapping using colour-coded blocks for structure, content, outcomes and assessment
- Co-pilot feedback on learning outcomes and activities
- Export options to Word documents, CSV, Moodle and Coursera formats
- Learning type tagging and analytics helping align activities with pedagogy from the start
- AI-generated content ideas and personas, tailored to online, blended or face to face delivery
Coursensu enables visual, iterative, and collaborative design. It helps teams co-create modules with a clear structure from the start, reducing back-and-forth later and making key pedagogical decisions visible early on.
Putting AI to work in learning design
Tim Hall, Senior Manager: Product Innovation, gave a short demo of the AI-integrated tools within Coursensu, showing how course outlines and activities can be generated and customised with prompts. For example, by describing a module on “playful learning,” the system created a course structure, suggested activities, and offered variations tailored to online delivery.
Standout insights:
- You can define your learner audience and format (online, hybrid, etc.), and Coursensu adjusts its suggestions accordingly.
- The tool supports creating templates, which could be reused across MOOCs or postgraduate modules, supporting consistency at scale.
These integrated AI features could drastically reduce the time it takes to get from “course idea” to a structured draft, helping teams accelerate design work while still leaving room for academic input and refinement.
Demo 2: Ben discussed an approach for designing assessments that work in the age of AI
Ben O’Hagan, Academic Editor, shared a model developed by Dr James Wood from Durham University for portfolio-based assessment, designed to make cheating with AI practically irrelevant. The multi-stage process includes:
- Early-stage planning and group annotation
- Draught submissions with AI, peer, and self-generated feedback
- A final essay accompanied by a process portfolio (including AI-use history)
- A short video explaining the student’s learning journey
Ben discussed how this model might be adapted for online learning and Canvas, including using tools like Studio, quizzes for self-assessment, and AI-feedback prompts via Noodle Factory or built-in AI.
As AI use becomes more embedded in student work, we need assessment strategies that embrace transparency, critical reflection, and process—not just product. This model does just that.
You can learn more about this approach in a presentation Dr Wood gave at the Birkbeck AI Summit, 22 July 2025:
Takeaways and next steps
Both demos highlighted a common theme: AI is changing not only how we design courses, but also how we design assessment. Taken together, these approaches suggest a shift towards more open, flexible, and resilient teaching practices. Rather than treating AI as a bolt-on or a threat, we can integrate it thoughtfully into both curriculum design and assessment design, supporting educators to work more efficiently and students to learn more authentically.
Behind the Screens gives colleagues a chance to share practical, hands-on work in a low-pressure space. It fosters a culture where challenges are shared early, prototypes are celebrated, and solutions are developed collaboratively.
This blog post was developed with the assistance of generative AI tools to support transcription summarisation of the meeting. Final edits and judgement were provided by a human author.
