The Future is Human – GenAI’s role within Communities of Practice (Part Two)
Part two of this series is going to cover the things we can do with Generative AI tooling to help us do the things we can do ourselves - but better
In the first part of this series, I wrote about the basic premise that just because we can technically do something with generative AI tooling doesn’t mean we have to or even should do it. If we automate away the tasks that bring happiness to us simply because they can be done a little bit quicker, then we can find ourselves without a reason for doing them in the first place. People join communities of practice to be around and share experiences with other people, and that’s the part that we should hold sacred.
Having said that, I am always an advocate for bringing new technologies into community spaces that provides an additive effect, and GenAI tooling is no different. There are lots of ways it can be used to enhance the human experience of participating in a community of practice, so I thought I would try to list some of them out. Then once I got started, I realised that there were good reasons to try and use tooling but also good reasons not to. So I thought I would share both, giving you the opportunity to decide whether or not it will work for you…
Doing heavy-lifting tasks
Summarising
Do the thing: I suck at taking minutes. I forget actions, I get excited about ideas and then realise I’ve been taking notes and not listening for five minutes and now everyone’s staring at me to say something, and sometimes I just don’t get or understand the point someone’s making. Beyond a tool like ChatGPT, there’s lots of integrated use of AI that can be used in parallel. Online sessions held over Zoom/Google Meet/Teams/whatever will usually have an option to record or transcribe the session, which when fed into a GenAI tool can remarkably quickly be summarised for future reference.
Also, depending on which tool you are using, you can feed them URLs, visualisations, studies, research, and ask for a summary that’s easy to digest/professional/long/short/bullet points/in any shape or form you want, and it will do it within seconds. Magic!
Don’t do the thing: Research has shown time and again that taking notes helps embed knowledge in our memories, and in particular writing notes by hand embeds it even more deeply. Also – sketchnotes! Sketchnotes are sick! My attempts are really terrible, but if there’s someone in your community who is good at them, they are such a special and treasured artifact. When someone has put love and attention into something, don’t let our immediate thought be how something else could do it faster. Instead, try to use it to augment instead of replace – include the fact you are intending to take sketchnotes in your summarising prompts, so that it can tailor its outputs to suit your needs more effectively
Supporting research
Do the thing CAREFULLY: We know that Generative AI models have a reputation for making stuff up. That’s because they are making stuff up – I mean, the clues in the name…its an artificial intelligence that generates things. There is an illusion of familiarity because you can use it in a similar way to Google and you will often get a more reliable answer. The modern internet is a mess - we know its full of adverts, sponsored links, spam, or something that sounds convincing from one of those lovely corners of the internet that are spewing misinformation into the world for whatever malicious or incompetent reason. We don’t always trust a search engine’s results, because we are familiar with this problem.
We should treat GenAI with similar scepticism – it is too simple for me to say that if you ask it to do something, its trying to please you. Its not. It doesn’t think. It doesn’t know you are on the other side of its interface entering information, all it ‘knows’ (spoiler alert – it doesn’t ‘know’ anything) is the probability that one chunk of word is the most likely one to come next based on the input it was given.
This doesn’t mean that it’s of no use and shouldn’t be used to support research. It should. As I said earlier, it often provides a more reliable response than a search engine does, because its better at doing ‘research’ than you are. If you only read the top response in a search engine, you only have one source of truth – you are dependent on the search engine to give you the right answer. Whilst we trust the algorithms to get us the right information, as mentioned above its not always the case - if it gives you the wrong answer, then guess what – you’re wrong too! Your GenAI tool’s model will have been fed millions/billions/trillions of pages and will be providing a response based on what it thinks is statistically the most probable information. Sometimes it gets it wrong. Usually it gets it right. The best tools will provide sources and references for their responses. Use these tools to help bring you to the right answer, just don’t solely depend on them to give you the right answer.
Reviewing whiteboards and outputs
Definitely do the thing: One of my favourite type of community session is when we all find ourselves around a whiteboard, pens and post-its flying everywhere, with all sorts of opinions and ideas being shared. At the end of the session we stand back, satisfied with our achievements, we take a picture of our handiwork for future reference, and then we move on with our lives. Sometimes, we even do some of the things we wrote on the whiteboard!
It would not be unfair to say that even with the best intentions, these sessions generate a lot of waste if not followed up on effectively. We write up things that we immediately forget about and never think of again. We forget the context surrounding one point and apply our own personal context and biases to another. Even the most successful ones require a facilitator or volunteer to go away and type them up into a longer lasting/more useful format. Fortunately, this is something that GenAI tooling can help with. And it can do so in some really clever ways. You can hand it the photograph you took and ask it to interpret and list out the actions. You can ask it to summarise them. You can ask it to list them out in markdown or some other computer readable format so that it can be fed into something else. You can ask it to group the outcomes in ways you didn’t get time to in the session. You can ask it to provide insights on the content generated that you might not have thought of. You can even use it as a tool during the live session and ask it list obvious insights and then provide discussion points that are more obscure or abstract to consider. This is a really powerful application for any thought worker focussed community of practice, and once again demonstrates the value comes from augmenting human creativity rather than replacing it.
Session backlog inspiration
Do the thing: Its hard work to come up with topics and ideas for sessions. Inspiration often abandons us to the whims of tedium, and we can find ourselves bereft of ideas and enthusiasm. Design your prompts to tell your GenAI tool to act as an experienced community leader or facilitator in your community’s particular domain space, and ask it to recommend some different sessions or topics your community should cover. Then ask it to be more creative and adventurous with its responses. Then ask for twenty more. Then ask it for some controversial topics that you might not have thought of. Then ask it for some ideas that wont otherwise be related but might have some sort of tangential connection and interest. Then give it Emily Webber’s kick off canvas and ask for suggestions based around that structure. And when its done that, take it to the community members and say ‘which ones do we fancy?’
Don’t do the thing: Communities are only successful if their members are invested in them. If you hand over the decisions on what the community will cover or focus on in their time together, we very quickly destroy the investment our people have in it. It becomes boring and formulaic, and devalues the experience of being part of it. Use the outputs above to prompt and to spark inspiration – don’t delegate authority for your community’s to a tool…….its a tool.
I think I’ve got one more of these in me for now, so next time I’ll write about using GenAI tooling for the things that we can’t do, and how as leaders we can use it to provide a better and broader experience to our members. In the meantime, I’d like to share this blogpost which goes into a lot more detail on how GenAI can be used in learning spaces – its very good, and well worth a read! ChatGPT Don't Care: AIs in Communities of Practice | Medium
Thanks so much for reading, if you’ve enjoyed this post I’d really appreciate it if you could share it - alternatively you could always buy me a coffee :)