Good Practice Highlights, The University of Sheffield

Making Sense of Conference Notes using AI

When attending events and conferences, notes are often made. You may have a preferred way to capture your thoughts. Over the course of an intensive day or few, these thoughts and observations can become a wide-ranging record.

Coming back to these a day or week later in order to make sense and share these with colleagues can often seem a large amount of work.

Find out one way of using AI to create digestable and useful outputs for yourself and to share with colleages.

James Little

Digital Learning Advisor, the Digital Learning Team

I attended DigiFest 24 - An education and technology conference in Birmingham, UK, for the first time. The conference was over two days, had a wide variety of sessions, talks and keynotes and was geared up to iknspire thoughtfulness over current educational practice issues as well as wider future challenges affecting us all.

For each session attended I made notes within Google Keep and also a photo to capture the atmosphere. This was to record not only a thread of each session, but also enable interpretation and consolidation of ideas and information that was being presented. Personally, this is how I ensure to capture areas of interest and order for future use ideas and concepts.

After two days I had 9 sessions worth of notes. In order to make a 10 minute team presentation on the conference I decided to try summerising the notes using Google Gemini, rather than doing this manually.

First, I fed all the notes into Gemini by copying and pasting content in one document at a time, and using the prompt for each session title, followed by the notes:

"Summerise the following session: Understanding the experience of students as avatars by Gary Burnett"

This produced a summary, split into main ideas; key points and action items.

I then briefly went through each of the notes I had written and checked that the summary was a true reflection of them and that it also made contextual sense. In each case it was accurate. Where relevant I added some further action points.

I combined all the AI generated summaries into one document.

Secondly, I then asked Gemini to summerise all of this into an overall summasry with the prompt:

"Please summerise the digifest conference 2024 from the following notes: ".

This produced a short summary of both days of the conference. I checked that this overall summary was a true reflection of my observations and thoughts of the conference, and in this case it was.

This process enabled me to synthesise notes into more concise summaries, provide an overall summary and do so accurately. This took far less time (45-60mins) than doing this manually. I then produced the following slide deck to share back with colleagues.

Observations

Checked own memories and understanding from reading notes and Gemini summary.
Gemini enabled a clearer understanding and recall than my own notes (!)
Checked for accuracy.
This all took (including this slide deck less than 2 hours). May have taken half a day to summerise notes previously and make presentation.

To note

Still could add own observations/final summary.

Enabled a test on colleagues to use AI meditated/summarised content - and revealsed:
- Last 5 bullet points is the only summerised session content which has been directly typed by me (apart from this presentation)
Still worth checking for sense/context - but can easily pick out themes and points quicker and potentially more accuratly than own.

Workflow Summary