Knowledge Summit Dublin

So, Monday and Tuesday of this week were devoted to the first edition of the Knowledge Summit in Dublin, with a strong AI theme, a pattern we saw last year at KM World.  It was held at Trinity College Dublin, and I could visit the old library, pictured to the left, along with the Book of Kells.  I was, for the first time, doing a joint keynote with my daughter with my daughter on day two, along with being a part of the closing panel.  Usually, I’m not too fond of joint keynotes, but this one was different.  Ellie took an anthropological perspective, questioning the issue of agency in machines, the need to realise the embodied nature of knowledge and the role of narrative, along with a few other things, and I picked that up with some broader implications and a presentation of the EU Field Guide recommendations in so far as they pertain to resilience and knowledge management.  The whole thing benefited from sitting through the previous day’s events to tune into the audience.

To make one point up front – this was one of the best-organised events I have attended in years (and I go to a lot).  Barry and his family and an army of volunteers left nothing to chance, and there were a thousand small things that were designed to make people welcome and engaged.  The event adopted a split format of 50% content and 50% discussion, although in most cases, that ended up as extended questions – but that is no bad thing, I am not sure the ‘flip’ was as much as intended.  We had more latitude in the keynote sessions, but we all respected the idea of that balance.  Trinity was a great location, and it was good to explore the town each evening, although I drank far too much Guinness.

For those who have been around in KM for three decades (and there were a few of us at this event, it was like slipping back in time to the turn of the Century), the regular rise and fall of KM, with each new surge failing to learn the lessons of the last fall(s) is both depressing and a little ironic given the learning promised by KM consultants.  The current interest in AI drives this latest ‘rise, bordering at times on obsession.  But the same mistakes are being made: talking about capturing tacit knowledge, the obsession with text and conventional information,  the desire to fit human knowledge exchange and creation into technology-based boxes and so on.

I picked up on various themes, which I list below, and this post talks more generally about knowledge mapping with links to other material.

1. The fact that it isn’t intelligent and, to my mind, will never be:  I wrote a series of blog posts on this subject, of which this is the first.  Scroll or use the category search to see the others.
2. Polani famously said that no explicit knowledge can exist without a tacit component, and I also expanded on his work to create one of seven rules of knowledge management, this one being we always know more than we can say, we will always say more than we can write down.   The requirement of everything to be explicit or tokenised is a limiting factor in making tacit knowledge explicit and on AI in general, which is dependent on inductive reasoning (the earlier referenced series covers this)
3. The danger is that we could lose our abductive capacity if we don’t use it
4. The isolation of humans from each other by technology and the severe dangers of that – especially as it is a commercial motivation behind the isolation
5. The social nature of human knowledge, Clark’s idea of scaffolding through narrative and making decisions in small social groups (more links on that in the future)

More generally, I pointed out the irony of the volume of information being generated by technology being used as the reason to create more technology to cope with the information glut.   Linked to that is the simple fact that most of the talks used ‘knowledge’ when they were talking about ‘information’/

I think it was videoed in which case it may be available – and having Ellie start the thing helped considerably.

Other reflections on the conference – I am not sure the flipping worked. People with good content didn’t have enough time to deliver, and the conversation became questions rather than some form of mediation.   It would probably have been better to run with two/three substantial keynotes and a few more panels with people working in groups in more extended mediated discussions.  Our triopticon method might be interesting to try, at least in part.  My only complaint was the promotion of NLP in one presentation coupled with mesmeric tricks – in the presentation style of an Irish Tony Robbins.  If it had just been the magic, it would have been a good break, but not the promotion of what I (and the reliable third-party sources) consider a manipulative and unethical pseudo-science.  Most people loved it, but I’ve seen too many people harmed by this sort of thing to stay silent.

But it was a great conference, one of the best I’ve been to, so when it comes up for 2025, please book – we may do some walking after it 🙂

Oh, and the opening picture is of the library. Books still matter, but also a bust of  Bacon, whose statement Knowledge is Power is one of the most misinterpreted sayings in the field.