NotebookLM summarizes my dissertation
I finally tried out Google’s newly-announced NotebookLM generative AI application. It provides a set of LLM-powered tools to summarize documents. I fed it my dissertation, and am surprised at how useful the output would be.
The most impressive tool creates a podcast episode, complete with dual hosts in conversation about the document. First - these are AI-generated hosts. Synthetic voices, speaking for synthetic hosts. And holy moly is it effective. Second - although I’d initially thought the conversational summary would be a dumb gimmick, it is surprisingly powerful.
This audio was computer-generated. Not just text-to-speech, but software summarizing a document and creating two virtual podcast hosts to “discuss” it:
(the audio file won’t play properly inside the AI-Generated Content callout block, but this audio file was generated by NotebookLM:)
Seriously - if you aren’t one of the half-dozen people on the planet who’s read the thing, but are curious about what I spent 6 years of my life working on, this’ll do the trick. Give it 12 minutes and you’re done.
And, dang. Sure, it misses much of the dissertation. For people who care about the details and nuance, they’ll read the thing. For everyone else? This isn’t a bad overview.
NotebookLM has a “Briefing doc” tool, which will produce a summary of a document to help with note-taking.
Of note:
- It doesn’t like my name - switching my first and last names and dropping the ‘A from my first-becomes-last name.
- Sure. It’s a summary. Did it capture all of the most important parts? Not really. But would it help someone get an overview of what to look for before they choose the parts of the document to read more closely? Sure.
Here’s the Briefing Doc for my dissertation, verbatim as generated:
🤖 - AI-Generated Content via Google NotebookLM
Briefing Doc: Integrating HCI and SoTL
Source: Excerpts from “ucalgary_2023_norman_darcy.pdf” (Darcy, 2023)
Author: Norman Darcy
Date: 2023
Main Theme: This dissertation explores the integration of Human-Computer Interaction (HCI) and the Scholarship of Teaching and Learning (SoTL) to analyse and enhance teaching practices, particularly in the context of online and blended learning environments.
Most Important Ideas/Facts:
- HCI and SoTL: A Powerful Synergy: The dissertation argues that HCI methods, traditionally used in video game design and analysis, offer valuable tools for understanding the complexities of teaching and learning.
- Games as Models for Learning: Darcy draws parallels between video game design principles and course design. He suggests that concepts like rules, feedback loops, and player agency can be applied to create engaging and effective learning experiences.
- The COVID Pivot: The COVID-19 pandemic forced a rapid shift to online learning, highlighting the need for innovative teaching approaches. The dissertation examines the challenges and opportunities presented by this transition.
- Visualising Course Design: Darcy employs visualizations inspired by video game design to represent online course structures, activities, and student engagement. This approach aims to provide instructors with insights into their course design and identify areas for improvement.
- A Framework for Integration: The dissertation proposes a framework that integrates HCI and SoTL concepts, providing a structured approach to analysing and enhancing teaching practices.
Key Quotes:
- HCI and SoTL: “Can HCI and video game methods be adapted to analyse classroom teaching practices?” (Darcy, 2023, p. 220)
- Games as Models: “These components of the classical game model are generalized, and are not directly connected to specific forms of implementation. They apply as well to the board game Monopoly (Magie & Darrow, 1935) as they do to Pac-Man (Iwatani, 1980) or Elden Ring (Miyazaki & Tanimura, 2022). They also describe aspects of course-based teaching and learning…” (Darcy, 2023, p. 102)
- Visualisation: “These visualizations of course designs and experiences by participants in various roles are similar to what might be produced to document a video game and its component levels, missions, quests, and challenges.” (Darcy, 2023, p. 178)
- COVID Pivot: “Examining COVID’s effects on Teaching and Learning” (Darcy, 2023, p. 130)
Research Projects:
The dissertation presents several research projects, including:
- Explorations: Teaching, Technology, and Performance: This project explores the use of technologies like computer vision and robotics to provide real-time feedback and support reflective practice in learning spaces.
- Teaching and Games: Connecting Performance-Centred Research Disciplines: This project examines how HCI methods, particularly those used in video game research, can be applied to evaluate and improve teaching and learning.
- The COVID Pivot: Documenting Change: This project analyses the experiences of instructors and students during the rapid transition to online learning brought about by the COVID-19 pandemic.
- Design Study: Exploring Online Teaching Performance: This design study investigates how instructors design and deliver online courses, using visualizations and concepts borrowed from video game design to represent their approaches.
Conclusions:
- HCI and SoTL offer complementary perspectives that can enrich our understanding of teaching and learning.
- Game design principles can inform the design of engaging and effective learning experiences.
- Visualisations inspired by video game design can provide instructors with valuable insights into their course design.
Future Work:
- Further development and refinement of the proposed framework for integrating HCI and SoTL.
- Exploration of additional technologies and HCI methods that can enhance teaching and learning.
- Investigation of the long-term impact of the COVID-19 pandemic on teaching practices and student learning.
Overall Impression:
This dissertation provides a thought-provoking exploration of how HCI methods and concepts can be leveraged to analyse and enhance teaching practices. The author’s innovative use of visualizations and his focus on the practical challenges and opportunities presented by online learning make this work particularly relevant for educators navigating the evolving landscape of higher education.
I then asked it: “How can insights from Human-Computer Interaction (HCI) research be applied to the Scholarship of Teaching and Learning (SoTL)?”
🤖 - AI-Generated Content via Google NotebookLM
Integrating HCI insights into SoTL
The sources offer many ways in which HCI research, especially regarding video game analysis, can inform SoTL practices.
- **SoTL, while valuable for understanding pedagogy and student learning experiences, often relies on reflections and recollections from participants, potentially overlooking real-time classroom dynamics.**Integrating HCI research methods can provide a more objective perspective on what transpires during a classroom session, supplementing traditional SoTL methodologies.
- The sources highlight “convergence” between video games and learning environments, making video game research methods particularly relevant to contemporary SoTL. Modern teaching, especially with the integration of online platforms, shares similarities with the narrative structures, performance-based activities, and interactive environments found in video games.
- HCI, with its focus on user interaction and experience with technology, offers a complementary perspective to SoTL’s focus on pedagogy and learning. This is especially important considering the increasing use of technology in education. For instance, analysing online platforms through an HCI lens can reveal design elements impacting interaction and engagement, which can be used to inform SoTL research on effective online teaching practices.
Adapting Video Game Research Methods
The sources specifically suggest adapting these video game research methods for SoTL:
- Formal Analysis:Â Analysing teaching sessions through frameworks used for evaluating video games, such as MDA (mechanics, dynamics, aesthetics), can provide structured insights into classroom activities. This could involve examining the learning environment’s design, the flow of information and control during the session, and the aesthetic elements influencing engagement.
- Identifying “Game Mechanics”:Â Borrowing the concept of “game mechanics” from video game design can help to identify and understand the underlying structures and rules that shape both course design and real-time teaching practices.
- Visualisation Techniques:Â Employing visualisation tools common in video game analysis, such as timelines and interactive diagrams, can help researchers better represent and understand complex classroom interactions and course structures.
Applying HCI-SoTL Integration
The sources present concrete examples of how HCI research methods can be applied to SoTL:
- Real-time Feedback:Â The computer vision prototype, designed to analyse movement patterns in a learning space and provide real-time feedback, demonstrates how HCI can be used to create tools that enhance awareness and understanding of classroom dynamics.
- Reflective Practice:Â The research involving humanoid robots, digital media recordings, and synthetic media highlights HCI’s potential for developing tools that support teachers in reflecting on and improving their practice.
- Course Design Analysis:Â Applying Fullerton’s Game Design Model to analyse interviews with award-winning online instructors proved effective in revealing patterns and insights into their course designs. This suggests that video game design frameworks can provide a valuable structure for analysing and understanding pedagogical approaches.
In conclusion, integrating HCI research, particularly those focusing on video game design and analysis, presents a significant opportunity for enhancing SoTL practices. By incorporating HCI’s emphasis on technology, interaction, and design, SoTL research can develop a more comprehensive understanding of modern, technology-mediated teaching and learning experiences.
The generated response also had footnote links to specific passages in the dissertation, and they seemed to be pretty accurate. This would be a useful tool to guide a closer reading. I wonder how much of my initial reaction is due to the quality of the response, or that it is so convincingly talking about my work in a way that is flattering. Of COURSE it’s a good tool - look at all the nice things it’s said about my dissertation! That podcast conversation has the potential to be simultaneously useful, and ego-inflating? And also missing some of the nuances - but that’s what podcasts with human hosts do, too…
I continue to be conflicted about the use of generative AI. One the one hand, the trope of “this is the worst that AI will ever be” holds true - it’s getting better and more capable every month/week. AND it is ethically questionable, resource-intensive, and actively pulling attention away from other important and not-just-invented-by-Silicon-Valley-VC-bros things (like Open Education, students as partners, experiential learning, decolonization, etc etc etc…).
And, for people who use this and come to depend on it - it’s from Google, so assume that it will have a limited shelf life before it gets sunsetted.
Comments powered by Talkyard.