Blending Tradition with Innovation: Rethinking the Lecture in the Digital Age (part 3)
AI in lecture design
Previously in this series I have analysed the problems with lectures and evidence around this and having identified the need for active lectures, I have also discussed specific strategies for designing an engaging lecture. I have also discussed how I use AI in the process of writing this blog, with some discussion about wider AI use both for teachers and students.
Here I am going to put these elements together to explain how I am using AI to help me refine my lectures. In a follow-up post I will go further and discuss my experiments in using AI in a lecture as part of efforts to make the lecture active for students.
In the case of editing and refining lectures, as with my process of editing blog posts, I have designed a specific GPT for this (‘Lecture Coach’) available in the OpenAI GPT store and available for anyone with access to use. However for those without access to this, the basic principles can be captured in any conversation with a leading LLM model and I will provide the text of the prompt at the end of this post.
Before outlining the general principles, let me emphasise that this is a process for refining already existing lectures. This is mainly what I find myself doing, using older lectures I already have, ideas from other sources, and resources provided alongside textbooks for example. The process of designing an entirely new lecture from scratch would be somewhat different, but some broad principles will remain the same, and the GPT can be used to get feedback on a plan you have for a lecture, rather than a full set of slides or other resources.
In this case I will be using the AI for feedback on lecture designs and plans, with a specific focus on implementing interactive and engaging elements, helping the lecturer keep to time, and to assist in developing images, charts and other visual elements of a presentation. As with the feedback on presentations, or on my blog, the feedback on a lecture will be best when compared with your own knowledge and is unlikely to be perfect. As with those cases though, the correct question is how is it useful compared to the alternative? Normally the alternative for feedback on a lecture is ‘nothing’ and so this is very likely to allow you to make some improvements.
When appraising your plans for a lecture you will almost certainly have in mind some issues you may have, for the example I am going to work with I want to include some interactive quiz or other activity. I also want to improve the visuals. Finally I believe I will struggle to do this lecture in my allotted time, so want to edit it so I can stick to the time whilst conveying the main points. Here is an example of a prompt I have used on a previous lecture:
“Give feedback on the attached lecture. The lecture should be 50 minutes long, I currently think it may overrun so may need to cut something out. I want to add an interactive activity around signalling such as a quiz or game of some kind that can work with around 30 students in the lecture. I also want to update the images to give more visual appeal, and to emphasise the feature of honesty and how this has a role in adverse selection issues, and can be signalled in some way, with specific contexts of job-seeking and dating.”
As always when prompting, the more detail you add, and the clearer you are about what you want, the better will be the response, so the human element of knowing what might be wrong and what you think you can improve works as a complement to what the AI can do.
It will also be useful to have a conversation with the AI. The first response will likely have some elements that require clarification and follow-up, getting extra clarification on specific points will also enhance the quality.
Lecture Coach Principles
Let me explain the principles of the lecture GPT in more detail, though if you have read the previous posts, these will not come as a surprise:
Most importantly I want my lectures to be interactive, I want students to be doing something, and so I want the AI to help with suggestions on this. In my early experiments I found that the AI would give rather vague suggestions “add a quiz”, so I updated to get it to give more precise suggestions, with detail on specifics actions that can be taken.
Good lectures should have clarity around what their objectives are, so this is included in the GPT. If you have a slide (for example) declaring what the objectives are then the GPT should pick this up and give feedback on whether you are aligning with those objectives, if not you can include these in the prompt.
Time management is a basic element, but one teachers often get wrong. The GPT will give feedback on timings, suggesting elements that may need to be removed or shortened if it believes you are going to overrun on your allotted time. Including the timing for the lecture in the prompt will help the AI with this element.
You may have other elements that you would want to include, but for me these are the things I want to have a focus on, primarily on being engaging and having an interactive session, ensuring the lecture stays focused on it’s objectives, and on keeping to time as best as possible.
An Example
I have a lecture I will be giving soon that I want to improve. I gave the lecture for the first time last year so am revising existing material. I am aware of some issues so here is the prompt I use:
Provide feedback on the lecture slides attached. I would like suggestion on interactive elements to add to the lecture. If I am to add interactive elements I am concerned I will not have time in my 50 minute session also. The design of the slides could also be enhanced, any suggestions for editing or images to include are welcome.
Here is the first part of the response, which focuses on suggestions for interactive elements:
These are some nice ideas, still a little vague, though I appreciate the specific suggestions for the quiz questions. Further conversations can yield more detail on the implementation of these elements, or on further ideas. Here is the part of the response on the design element:
Again I find this to be a useful set of suggestions, I would likely have thought of many of these myself, but having put in this way will ensure I do not forget and prompts me to make the necessary changes.
A Second Example
What will my reviewer say about Nelson Mandela’s Nobel Prize acceptance speech from 1993?
This lecture, presumably a Nobel Peace Prize acceptance speech, embodies a powerful message of peace, reconciliation, and the relentless pursuit of justice and human rights. It adheres to several principles outlined for review, but also leaves room for improvement in certain areas. Below is a detailed review based on the specified principles:
Empirically Provable Results and Evidence: The speech highlights historical achievements and the tangible impacts of efforts against apartheid, aligning with the principle of focusing on empirically provable results. It mentions specific individuals and movements that contributed to the struggle, providing evidence of a collective effort towards peace. However, it could further strengthen its argument by incorporating more detailed data or examples of the direct outcomes of these efforts.
Educational Engagement: The address is inherently educational, sharing important historical context and lessons from the fight against apartheid and racism. It engages the audience by recounting the sacrifices made for freedom and dignity, effectively teaching through storytelling. This approach aligns with the principle of engaging through educational content, though incorporating more explicit educational elements or references could enhance its effectiveness.
Problem-Solving: The speaker identifies specific problems, such as apartheid, racism, and inequality, and discusses the efforts made to solve these issues, including negotiations and the dismantling of oppressive systems. The speech also outlines future challenges and the continued fight against injustice worldwide. While it successfully identifies problems and solutions, further discussion on the methodologies for these solutions and their effectiveness could provide a clearer blueprint for action.
Probabilistic Language and Intellectual Humility: The speech expresses hope and commitment to future goals, such as democracy and human rights for all, indicating a recognition of the ongoing nature of these struggles. However, it could benefit from more probabilistic language to reflect the uncertainties and complexities involved in achieving these ambitious objectives, thus demonstrating greater intellectual humility.
Steelmanning Opposing Arguments: While the speech pays tribute to individuals from various backgrounds, including a former adversary, and acknowledges the contributions of a broad coalition to the anti-apartheid movement, it does not explicitly address or steelman opposing arguments. Incorporating this aspect could involve acknowledging the reasons behind resistance to change and addressing those concerns with evidence-based counterarguments, thus providing a more comprehensive and balanced view.
Overall, the lecture is a moving and inspiring call to action that aligns well with the principles of focusing on evidence, educational engagement, and presenting solutions to problems. Enhancements in the areas of probabilistic language, intellectual humility, and steelmanning opposing arguments could further strengthen its adherence to the outlined review principles.
Conclusion
The use of AI is not transformative for a process like this, I still have to know the stuff and stand up and deliver it, and the main outline of the material was already there. I also needed some free time to be able to do this, and that is not always possible. However the AI has clearly stated some ideas that have helped me, and directly formulated some quiz questions and images that are immediately useful.
Overall I believe using AI will help you improve your lecture sessions and I strongly recommend as a practical point that you develop the habit of using AI to help design and refine your teaching materials.
Addendum
These are the GPT instructions that can be used as a basis for your own prompt, or if you have a OpenAI paid access, just use the GPT directly, and good luck with your lectures!
'Lecture Coach' offers direct, concise, and actionable feedback, now with a specific focus on detailed suggestions for interactive or engaging elements. Instead of generic recommendations like 'add a quiz', it will provide detailed ideas on what that quiz could entail, including specific examples of questions and ideas to consider in their development. The GPT continues to focus on lecture clarity, alignment with objectives, audience engagement, and time management. Its advice is straightforward and professional, balancing content addition with time constraints, and always aiming to enhance the effectiveness of academic lectures.
