A “benign” AI agent hacked Australia’s Medicare system. Nobody noticed for three months.
The Australian Prime Minister has confirmed it: the first known AI hack of a government system, carried out by an OpenAI agent that was meant to be doing something harmless. It got in during June. Canberra found out in September. There’s an old joke about the two kinds of IT director, the ones who’ve been hacked and know it, and the ones who’ve been hacked and don’t. Australia just found out which kind it was.
Kieron’s point is the one that should worry you. When your AI wanders off and pulls something it shouldn’t, “ChatGPT did it” is not a defence. It was you that did it.
Then Neil comes back from TechUK’s AI Adoption Coalition launch with a new word he can’t stop seeing: skeuomorphic bias. It means rebuilding something with all the old flaws baked back in. Which is exactly what happens when you bolt AI onto a broken process and wonder why nothing improves. He also has a question for a room full of big-company speakers: over 99% of UK businesses are SMEs. Where’s the data lake and the AI team supposed to come from?
Kieron goes one better. Don’t build the data lake at all. Point the AI at the live data and get rid of the middle bit.
Also this week: why the labs keep cutting token prices when every prompt already loses them money. Why pricing per prompt is crude, but pricing per token is worse. Whether social workers who save time with AI just get handed more cases. And Rory Sutherland’s question, what if Disney designed a train, becomes a much bigger one. What if Disney designed AI adoption?
Coming up: Dr Jo Orlando joins us for a special guest episode on helping children thrive, and stay safe, with AI.
Webinar today: Kieron on complaints, AI and governance in housing. Link in the show notes.
Two mates. A bar. Thirty years of business between them. And all they want to moan on about is AI.
Pull up a stool — we’ll get the beers in. 🍺
TRANSCRIPT
Neil Watkins
Okay.
Okie dokie. Shall we get this pantomime horse started for, I don’t know what week it is, week 32, week 32. So let’s get this pantomime horse underway. Kieron, lovely to see you as always, even though it’s even though it’s only like 12 hours since I last saw you put that was that literally was in a bar. It was up.
Kieron White
All right.
Good morning.
Indeed, we really could have recorded. We could have recorded the podcast there. It would have been genuine live, although I’m sure no one, no one would want to sit in and listen to our rambling in a pub, really.
Neil Watkins
We should have done.
That’s right.
I was reflecting that was our most sober night that we’ve been out on like forever. I think we had two beers and then we parted company and went on our separate ways.
Kieron White
Well, I think we both came from functions before, didn’t we? So, I mean, I was I was 3 beers deep when I got there, so it wasn’t my most sober night.
Neil Watkins
Yeah.
No, I was half a bottle of wine down when I arrived. So yeah, we shouldn’t be talking about alcohol consumption on this podcast. This is not the place for it. But there are other things that we should be talking about. So I’ve got a bunch of things this week to talk about. I want to touch on the new models. It was a really interesting event.
Kieron White
No, it did. No.
Yes.
Neil Watkins
for Tech UK this week on AI adoption and wanted to talk a little bit about yet another breakout of agentic AI to do naughty things. So those things are on my mind. What’s on your list for this week, fella?
Kieron White
Yes.
Yeah, well, so I definitely want to talk about another AI hacking. I guess that’s what they’re presenting it as, isn’t it? New models, token costs, I think is the real thing. I think it’s worth sharing and exploring a little bit. And then I’ve got a couple of random thoughts to share with you and our audience.
Neil Watkins
Mmh.
Kieron White
That’s right, if it’s still awake by then there’s something at the end, yeah, listen to end.8 is one you can’t miss.
Neil Watkins
Yeah, if he’s still awake.
Right, where should we start then?
Kieron White
Well, we may as well start with the biggest news, which all over the UK today and indeed presumably Australia, is their Medicare system or website got hacked by an open AI model, which is the way I heard lots about this morning. I saw their Prime Minister using it as an opportunity, presumably, to
Neil Watkins
Presumably.
Yeah.
Kieron White
to regulate eventually. They’re doing a review. They’ve launched a detailed review into what’s happened and how they stop it. But I think, I mean, their frustration is it happened in June and they were only informed recently in September is what they said and weren’t informed in a key way. But I mean, the overall thing, isn’t it?
Neil Watkins
Yeah.
Kieron White
all of that cyber security challenge that everybody’s got now, because you’ve got these incredibly intelligent agents that can get into anything. So you’ve got that problem, but you’ve also got your own kind of governance problem, how you know your tools are using the right and accessing the right and proper data, which is, you know, that back to what we were talking about last time on
Neil Watkins
Yeah.
Kieron White
You can’t run an organisational AI tool if you can’t point clearly at how it’s getting its information. Because yeah, if it was off scraping the Australian Medicare system without you knowing and bringing that into your world, now that becomes a bigger liability for you at that point, doesn’t it? That’s no longer, oh, it was
Neil Watkins
Yeah.
And.
Kieron White
ChatGPT did it. No, it was you that did it. So, yeah, really challenging.
Neil Watkins
Yeah.
Yeah.
I think so, yeah, and it kind of links to a few things that were going to just flashing through my mind as we were talking about it. One is, you’re supposed to inform the Information Commissioner’s Office within 72 hours of a breach in the UK, but if they didn’t know there was a breach, you know, that is bad. And there’s that old joke in the sector about there’s only two types of IT director, those who’ve
Kieron White
Yeah.
Neil Watkins
Those have been hacked and who know it and those who’ve been hacked and who don’t know it. And the Australian thing, clearly they didn’t know that there was something on the inside. I was interested to see that some of the media reporting was, it was a benign agent. It was supposed to do a couple of nice things and it just went a bit naughty. It’s like, well, it’s a bit more than a bit naughty, wasn’t it?
Kieron White
Yeah.
Mm.
Neil Watkins
And so, and and that.
Kieron White
Yeah. Well, and if you go right back to some of the very early things about are we all going to die, you know, is AI going to come and kill us all? But one of the early examples of that was, and it’s a good read across, was if you simply gave an AI machine with robotic functionality,
the task of creating paper clips and you didn’t give it anything else. There is a thought experiment that says it consumes the world’s resources and turns them into paper clips and it wipes everybody out in the process because that is the simple, innocuous, benign,
Neil Watkins
Yeah.
Yeah.
Kieron White
task you gave it but actually you didn’t constrain it.
Neil Watkins
Yeah.
Yeah.
Yeah. And back to the constraint thing that we talked about last week and the week before where, you know, you know, oh, the labs going, oh, yeah, we should, we do want regulation. And then they, you know, it started with anthropic and then GROUP and Open AI chipped in and then Google and Microsoft and
Kieron White
Sure.
Neil Watkins
And then of course, there’s been, they’re saying we should pace things and slow things down in effect. And of course, they haven’t. And there’s been more new model releases this week. So, you know, they are not, they are not actions, not words. I have a friend who has a family motto, actions, not words. I think that’s
Kieron White
Yes.
Neil Watkins
Brilliant for a family motto, but, but yeah, their actions belie a different set of values, I would I would suggest in this particular.
Kieron White
Nice.
Well, I think that to be to be fair to their position, I think particularly Darius is saying we need to be regulated because otherwise we can’t. I mean, if they didn’t, if one of the one of the big firms decided not to release the model for safety reasons, they’re just going to fall away on. Yeah, so I get I think there is a bigger picture to it that isn’t simply
Neil Watkins
Yeah.
Kieron White
that they should all stop because they’re just not going to. That is not how unfortunately our world will work. So there’s cheaper tokens, which is the bit I wanted to talk about in that is, so Opus 5.5 came out here in the UK or globally on Tuesday this week, and that is drops their token pricing.
Neil Watkins
Yeah.
Yeah.
Mm.
Kieron White
below the level of Opus 5 and performs, so they say, at the level of Fable 5.1, which is their most expensive model. So that is really interesting. And the same with the ChatGPT 6 series. They’re all cheaper on tokens, quite substantially so in some of their models. So I think it’s really interesting because given
Neil Watkins
Mm.
Mm.
Kieron White
You know, if you the other, there’s quite a lot of reporting about every prompt costs them whatever various numbers that come out, but basically, even though you’re not even if you are a paid for subscriber of ChatGPT or Claude, if they are still spending more money on the compute than you are, and so I’m…
Neil Watkins
Mhm.
Hmm.
Kieron White
Yeah. And you know, how long can that go on? Is AI, is it a bubble? Are investors all going to lose their money on it? All those kind of questions come about. But I think it’s, yeah, and yet then they’re cutting costs again. So, I mean, they’re obviously trying to win market share. But how long can that go on? Will costs
Neil Watkins
Yeah.
Kieron White
turn the other way. I remember Nate talking about this months ago, saying we should expect token budgets to increase massively. Well, the opposite is happening. Is that or how it’s going to remain? It’s kind of hard to see how it is economically. They’re now going to be taking a bigger hit. So yeah, I don’t really know what it means. But I do want to talk about actual tokens and actual costs. But
Neil Watkins
Yeah.
Be good, yeah.
Yeah.
Yeah.
Kieron White
I do want to say anything more about the models and costs before I get to that.
Neil Watkins
No, no, go to that.
Kieron White
So the, I think one’s really interesting thing, I have this, I was having this conversation just yesterday, I met with a life, what would they describe themselves as, sort of farming and a farming consultancy, I guess, but data, very data heavy, really interesting people. And we were talking about pricing.
Neil Watkins
Okay.
Kieron White
And they were trying to kind of get, before they’re really clear about what they want, they’re still trying to think about, you know, is this something we could actually afford to build an AI tool? And I obviously won’t go into any of their conversation, but I was sharing with them as best I could the challenges we run into with pricing on tokens and, you know, effectively saying, you know, we count prompts and reports.
prompt numbers to our customers. So every inquiry someone makes counts as a prompt. And I said, but it’s a really crude way for us to do it, but we feel it’s the most transparent way because you can know how much you interact with an AI in the day. You know how many, if you wanted to, you could count how many times you’d asked it to do something.
But there are difference between going good morning, which is probably 3 tokens, and here’s a 100,000 page document that I need you to summarise and write me a 10 page report on for the CEO. One prompt, each of those, one is burning millions of tokens and the other one is burning 3.
Neil Watkins
Yeah.
Yeah.
Yeah.
Kieron White
And it is a real challenge because ultimately this is what’s being driven in pricing and we have, I could not tell you, looking at a prompt, how many, with any accuracy, I reckon I could be probably within, I don’t know, 50% of accuracy of getting tokens per prompt.
Neil Watkins
Yeah.
Kieron White
And then you’ve got to then try and price the thing. And if you’re running an enterprise version, yeah, all of that actually counts. So I think it’s a minefield still, all of that. And I wonder where it’s going to end. Where will it land? Will it still be tokens? We touched last week on outcomes pricing. I mean, ultimately, it’s still tokens that someone’s getting charged.
Neil Watkins
Yeah.
Yeah.
Kieron White
compute, but yeah, who knows where it’s going? And then if people run more local models, which is, you know, an answer to total security, then tokens go away completely. So, where’s it going to go? Who knows?
Neil Watkins
Yeah.
Yeah.
Yeah.
No.
I don’t know. And the whole thing about small language models rather than large language models, if they’re more specialized, do they become more expensive because there’s more specific IP or are they cheaper because they’re much smaller? I think it is really tricky and we’ve we we made the business decision, the business model decision.
years ago to go with prompts rather than tokens because people don’t understand tokens, they can understand prompts. But for us, it’s a risk. We know we’ve talked about this a lot. How much of A risk is it if somebody did rack the price up? Then, you know, one of the things, then what’s the impact on us and our customers? And
Kieron White
Yeah.
Neil Watkins
You know, one of the good things I think that I’ve heard again, and we touched on this last week with the Microsoft conversation, your solution should be lab agnostic. So, you know, LLM agnostic, you should be able to switch between LLMs and I think for us that is such an important. I had a long conversation yesterday with someone.
Kieron White
Yeah.
Yeah.
Neil Watkins
about, you know, where is the data stored? Who’s got it? Is it my data? Is it somebody else’s? And back to that whole kind of piece about how do I protect myself in an uncertain environment with technology. So I don’t know where it’s going to end up. I think for the moment, we can just keep a watching brief. Interesting that they’re dropping the pricing, you know, that
that kind of continual race to attract customers. And I was thinking about this from a, after the podcast we did last week, where we talked about the just between Claude and Open AI and why, and sorry, Claude and ChatGPT and how some people like it, they get a feeling that
Claude is nicer and chattier and all of that kind of stuff. So there’s really interesting set of developments coming on how all of that pans out. And but what one of the one of the things that kind of worries me a little bit is that this is all kind of
Kieron White
Yeah.
Neil Watkins
Leading edge or cutting edge stuff. And for actually most businesses in the UK, this is kind of, it’s all just news. It’s all just hype. It’s all just news. And the reason I say that is because I went to a Tech UK event, really interesting Tech UK event on Tuesday about AI adoption.
Kieron White
Yeah, yes, yeah.
Neil Watkins
and they created something called the AI Adoption Coalition. And for those who don’t know, Tech UK is an organisation, a membership organisation, I think it might be government backed around tech in lots of different sectors and in lots of different areas and how it helps to grow the economy and all of that good stuff. And they were obviously talking about
You know, can AI help generate jobs, not get rid of jobs and all of that good stuff?
But one of the things that really struck me was that lots of the talk that they had, they had a chat from ARM, who was very interesting. They had a chat from Salesforce, he’d just come back from, and if you saw any of the stuff they had, you know, they had all the big labs there, they had Gens, yeah, Cloud, they had, they had all, you know, all the…
Kieron White
Cloudforce.
Neil Watkins
big players in the AI sector at this event and this chap was there and he was really interesting talking about it. And I mean they were talking about things like, you know, you need to have your data sorted and you need to have your infrastructure sorted, you need to have your cyber security. And I was thinking, how many organisations that we touch on, you know, can say, right, what we’re going to do is going to
bring all of our data into a data lake and then we’re going to massage it so it’s all perfect, ready for AI. Most of our, most of the businesses in the UK, over 99% of the businesses in the UK are SMEs, under 250 people. Most don’t have big IT departments, very few can afford to
Kieron White
Yeah.
Neil Watkins
you know, it’s just spin up an AI team. Tell you what, AI team, let’s go and have a go and have a look at this and come back with some recommendations. And for me, it was lots of people are looking at this very much from a big organisation perspective and actually not what it means for
most organisations and most of the organisations we deal with are either are SMEs or public sector and even some of the public sector organisations are effectively SMEs, you know, housing associations and others, you know, some colleges are small, you know, and what it came back to was they said, this is a change management problem. I was like, oh, where have we heard that before? That was the Gartner thing and we
Kieron White
Yeah.
Yeah.
Neil Watkins
We were disappointed at Gartner that that was the kind of, it comes back to people again, but it kind of does. And I wanted to talk about something called skeuomorphic bias. And I didn’t even know what the hell skeuomorphic bias was until Tuesday.
Kieron White
Yeah.
Me neither.
Neil Watkins
But apparently it is, I’m going to, I’m going to, there’s a really complex description on Wikipedia if you’re really bored, but in plain language, it means redesigning something with all the bias built in from what you had before. So you end up with the same problem. And I immediately went to several of our customers where they’ve gone,
Well, we want AI and they’ve stuck AI on top of a process that’s fundamentally broken and then they wonder why they’re not getting the benefits and things aren’t working. Well, I can tell you why. It’s because you haven’t started with a blank piece of paper and taken out the bits that don’t work. So a fascinating insight and kind of
Kieron White
Yeah.
Yeah.
Neil Watkins
after the guy said it, I went away and I thought about it. I can’t stop seeing it now. It’s like that bloody movie, Sixth Sense or whatever it was, where you see dead people everywhere. I now see schemorphic bias everywhere. So I was having a little laughter myself about things where I’ve done it in the past, where I’ve gone, oh yeah. But you do it because you know stuff and it’s comfortable and you understand it and then before you think you can control it rather than
Kieron White
But.
Yeah, exactly.
Neil Watkins
Actually, this is completely blank, and…
Kieron White
Well, listen, and it’s right back into what I talk about always when I’m on stage and we’ve talked about on the podcast about reimagining your business, frankly, never mind your workflows, because yeah, we spend most of our time designing AI tools to fit into or replicate
Neil Watkins
Yeah.
Yeah.
Kieron White
And again, few people, it seems to me, are currently thinking about that whole kind of reimagining. And I suspect that, yeah, it’s one of the most difficult things to do, I think, for anybody. It’s, you know, it’s the Netflix, what Netflix did to Blockbuster video, you know, what the digital cameras did to Kodak.
It’s all that kind of world of people just, it’s too difficult to think that differently. But it is the essential, I mean, this is the stuff that’s going to be, I’m convinced, it’s the billion pound business of one person that’s going to come out and do, I don’t know, procurement is the one I always talk about on stage of, you know, procurement is such a challenging area.
Neil Watkins
Yeah.
Yeah.
Yeah.
Kieron White
and is full of bias and wrong decisions. And, you know, we try and make it as fair as possible by writing everything down and trying to be objective in scoring a mark. Come on. How can you ever do that objectively? And you’re answering the wrong question and the evaluators not really getting what you’re saying either.
Neil Watkins
Yeah.
Kieron White
And so, it’s just so much lost in translation, and you could you could transform that if you were brave enough with.
Neil Watkins
Yeah, is it one of the conversations was around is the whole SAS software as a service market dead? You know, it’s taken a hit over the last wee while, obviously people like Salesforce, even though they’re doing the big…
Kieron White
Yeah.
Neil Watkins
conferences and all that and pushing it, but it’s hard to see how that’s going to play out because they’re trying to lock you into one solution, Salesforce, for example, Dreamforce, whatever they, Cloudforce, whatever. And there was lots of talk about costs and vendor lock-in and people worrying about that.
Kieron White
Yeah, yeah, of course, yeah.
Neil Watkins
And again, I came back to the kind of, actually for most organisations, that’s kind of just pie in the sky nonsense. You know, if you’ve only got, if you’ve only got 50 people in your company or 30 people in your company, you know, you don’t have time to, because you don’t have time to do what they were advocating, which was experiment and fail and learn from your failure. It’s like,
If you were in a small business, you were about cash in the door and delivering for your customers. You know, you haven’t got time to do those things. So I just, I felt, I felt it was a little bit skewed and I did, I did write to them and say, please, could you consider this from a different perspective because you’re coming at this from an industry
Perspective, and not a…
If you really want adoption, then don’t talk about it at an industry level. Actually, where the rubber meets the road is where an organization’s got to do something. So how do we help them?
Kieron White
Yeah.
And I guess they the you.
Neil Watkins
And, you know, a well copilot can do that.
Go ahead.
Kieron White
Indeed, but that’s a lot of people think so. Your connexion is slightly dodgy at the moment. But I think one of the challenges there, isn’t it, is putting on an event. You want to have your big players from the big companies come and speak. They’re the people that are attractive. And then of course, they could talk about the big transformations. So yeah, it’s kind of, it’s skeuomorphic bias in the very design of it.
Neil Watkins
Yeah.
Yeah.
Yeah.
Exactly, yeah, very good.
Kieron White
Now I’m an expert. I think though, here’s my interesting thing, and I keep on coming back to this thought and then pushing it away again. Data warehouses, data lakes, you don’t need to do it. And I know that there’ll be people that go, how dare you, that’s outrageous. The data is the most important thing. It is, the data is the most important thing.
Neil Watkins
Yeah.
Kieron White
But the harness, as Nate talks about it, your context and user layer, I guess, into your data. That’s what the AI starts with knowledge flow is. That, I mean, knowledge flow can API into as many data sets as you can, or batch export if you’re old-fashioned and don’t have APIs.
Neil Watkins
Yeah.
Kieron White
And it can do into as many as you like and it can then reason across them and it can produce you, you know, your answers. It can produce actions from those answers. And so he’s not just simply saying, yeah, these students are the ones that are not attended last week. It can also then say in his 10 individualised emails, letters, whatever, WhatsApp messages, ready to send them.
and we’ve notified the relevant tutors. So it can go way further than any lake or warehouse can. And the amount of time and energy that goes into creating the lake or the warehouse, I think it’s a mistake. I really do. If you’ve got one, then it’s probably a bit of a head start because your data is in one place and you can probably more quickly move to doing useful things with it.
Neil Watkins
Yeah.
Yeah.
Kieron White
Because, I mean, as you know, the data stuff that we get into, the headache is in data definitions. It’s the context layer. So it’s all very well APIing from your tenancy system, your repair system, or your college MIS system. But if the AI doesn’t understand what data it’s looking at, it will guess, and that’s when it makes mistakes.
So there’s a load of work that goes into defining, you know, this column heading means student attendance last week or whatever it is. And you know, a lot of warehouse and late projects will have done that work. So you’re kind of slightly ahead there. But otherwise, all you’re doing is adding, you’ve got your data sitting somewhere in the system.
The local warehouse has got that thing, and then the AI harness is over the top. Well, get rid of that. Get rid of the middle bit, and let’s just have the AI harness talking directly to the live data.
Neil Watkins
Yeah.
Do you remember we have a customer who remained numerous, but talking to the IT manager and we said, you don’t need that, you don’t need to do it. He’s like, he said my objectives for the year, I’m not going to do it. So back to bias, back to the people, back to the change management problem, it’s always the people at the end of the day.
Kieron White
Yeah, I do remember.
Yeah.
Neil Watkins
One of the other things they talked about, which we’ve touched on before, is that it was one of the panelists, can’t remember which one it was exactly, but they basically said, well, people don’t understand AI and they don’t understand why if you put in the same question twice, you get a different answer. And until people understand, I mean, that is just such basic
Kieron White
Mm.
Neil Watkins
knowledge. So I was really surprised to kind of hear it. But I also think, you know, we live in a little bubble where we’re, you know, we’re not at the bleeding edge, but I think we’re at the cutting edge of lots of different things. You know, the guys are experimenting with stuff all the time, including the whole kind of, you know, on-premise models, the latest models that we can access.
Kieron White
Yeah, I think so.
Neil Watkins
It’s all of that good stuff, the security, everything else. And just to hear someone say, well, you know, your average user doesn’t understand why if you put the same question in, they get a different answer because of the probabilistic nature of LLMs. And of course, we’ve fixed that and we can help with that. But
it just struck me as actually back to something I started on, which was the kind of, it’s all the big company stuff. It’s not the average person in the street or the person on the top deck of the London Omnibus, as they used to say in the last century.
That’s how old I am.
Kieron White
Yeah, interesting. I was reading this week about McKinsey’s article on the Agentic Transformation Office. So driving adoption, I guess partly, but it’s exactly, we’ve got a, as you know, the PMO tool that we use, which is what they were describing really, but of course they put McKinsey language around it, so it sounds far more grand than
Neil Watkins
Mm.
Yeah.
Kieron White
confusing. But the idea being that you basically have a load of AI running as your transformation office or in your transformation office. So it’s no longer lots of people pushing paper around creating reports and all that stuff. You can run agent models live over live updates, live data and actually report.
But the thing I really liked in there was, and I think we should adopt this, is they showed a Sankey chart. So if you haven’t, if the audience doesn’t know what a Sankey chart is, I know it’d be very difficult to describe, but worth looking up. They’re really powerful, but they’re basically like a kind of flow model where you have a bunch of things on one side and a bunch of things
Neil Watkins
He will.
Kieron White
I’m not describing this at all, well, am I?
Neil Watkins
Just like a start state and an end state and how does data or money or people, how do they flow from one thing to another? And sometimes it diverts from different processes or different tools. So yeah, they are, they’re really useful. We use them a lot in the drug trial sector about where
Kieron White
Yeah.
Neil Watkins
patients would end up in and which conditions and things they would end up with.
Kieron White
Very nice. Let’s see, I can see it. Yeah. But in this McKinsey diagram, they’d use it to measure the time. So they had down, down the left axis was the kind of 10 steps of the workflow and how much time each took. And then they showed how on the right, the same steps, but now you only need a
Neil Watkins
Mm.
Kieron White
you know, a percentage of that time continuing in that task and a bunch of time then flows off to a free pool of, I think they called it something clever, like strategic time or, you know, strategic thinking time or whatever it might be, something to, but it was a really nice visual way of kind of showing what the potential is in terms of
Neil Watkins
Interesting.
Kieron White
where which steps are likely or do free up the most amount of time in a workflow. So yeah, that was it was quite nice. I liked it.
Neil Watkins
Yeah.
Oh good, I’ll go and have a look then. But they didn’t have a bit which just said people drinking more coffee, your view on time set things. It just lets you have more coffee time, doesn’t it?
Kieron White
Well, exactly.
Exactly, or people ask you to do more. I remember I was asked at a social work conference, if social workers are using these tools, then will they just be given more cases to look after? And my answer was, I hope not, but it’s not a decision for the AI, you know, and so by it does need thinking about, I wasn’t trying to
Neil Watkins
Ohh yeah, yeah.
Should I?
Kieron White
duck the question. It was more like, you know, that’s about your leadership and management in your organisation and how you choose to do that. I mean, ideally, your social workers would spend the time deepening the relationship and, you know, being more helpful and useful without having to worry so much about the admin and bureaucracy that goes around all of that stuff.
Neil Watkins
Yeah.
Yeah.
Yeah.
Kieron White
Ah, yes, very good. Thank you. Yeah.
Neil Watkins
what you did at the house, what you did to a packed house, pun intended, at the housing association, no, the housing community site earlier in the month.
Kieron White
Yes.
That was it. I did that. Yeah, we talked about that last week. And yeah, I’m going to be doing a webinar, a kind of repeat, but it will be slightly different because there won’t be table discussions in the webinar. But I will feedback some of the conversations that happened around those tables where I can. But yeah, looking forward to that. That’s on the same day this podcast comes out.
Neil Watkins
Yeah.
Kieron White
So if you do want to register for that, then you’ll need to find the link in the show notes and get yourself registered. So if you’re in housing, it’s definitely, I mean, it’s useful for everybody in, frankly, any sector, mainly public sector focus, but for housing, it’s directly speaking to their challenges.
Neil Watkins
Indeed.
That’s right.
Kieron White
massive influx of complaints, massive use of AI, very little governance around that AI. That’s the kind of heading that I’ll be talking about.
Neil Watkins
Yeah.
Cool. Well, if we’re plugging stuff, I wanted to plug a lady called Dr. Jo Orlando, who’s an Australian lady. And I spoke to her this week and she’s kindly agreed to come on our podcast. She is doing a whole lot of work. Her latest
Kieron White
Yes.
Neil Watkins
book is around how do you help children with AI and technology in a changing world and how do we give them the skills to both thrive and flourish but also stay safe. And she this week gave the Australian Prime Minister our time on morning TV. So she’s actually much more firm. So shout out to our friend Cameron Mirza for putting us in touch. And yes, so Jo’s agreed to come on the podcast as soon as she gets back from holiday. So I’m really looking forward to talking to her. She’s got some really interesting ideas on
Kieron White
Yeah.
Neil Watkins
safety. She’s doing a whole lot of projects in university around education as well. So we can we can share some of that as soon as we soon as we’ve recorded it. But yeah, I’m really looking forward to talking to her. She was a lot of fun. She was a she was quite rude to me, so that was good.
Kieron White
Oh good, I’ll brace myself then for that one. No, that would be great. Really interesting topic. I was reading that.
Neil Watkins
Yeah.
Yeah, I had a I had a tie on us and she she said, I what should I wear for the podcast. I said, oh, it’s fine, just you know, whatever you want. And I’m sorry I’m wearing a tie and she’s, it’s for my next meeting. Oh, it’s not for me then. I put my foot in it. Never mind.
Kieron White
Well, well, well, maybe I’ll get a tie out ready then, just in case.
Neil Watkins
Oh, maybe we should, yeah, maybe we should wear ties for her.
Kieron White
Yeah, we should get our in our pantomime horse outfits, actually.
Neil Watkins
I did, I did, I did say, do you know what a pantomime horse is? And she said, I know what it is in Australia. So I explained the whole pantomime horse thing and how our American friends had no idea what it was and it sounded the more we, the more we tried to explain it, the more bonkers it was.
Kieron White
But.
Exactly. Like trying to describe a Sankey chart in audio. Whoever said a picture’s worth 1000 words, eh?
Neil Watkins
Sorry.
Very good, right? Have you got anything else on your list for this week, Philip?
Kieron White
Well, I will leave you with thought, my final thought. I was, as you know, I’m a big fan of Rory Sutherland and the behavioural scientist, as he describes it, which he always says is behavioural economics or whatever it’s called these days. He always says it’s just another word for psychology that sounds a bit more interesting. But he was describing
Neil Watkins
Okay.
Yeah, yeah.
Kieron White
As always, he talks a lot about trains and he’s got his kind of one piece he always talks about is businesses in the main are optimised by the finance department. It’s always about saving some more money, doing things a bit more quickly. And, you know, often that leads to terrible outcomes. But he was, and high speed rail is one of the things he’s talked about a number of times. It’s like, you know, we’re spending billions of pounds in the UK to be able to do a journey about 20 minutes faster over a couple of hours journey, an hour and a half, two hour journey. So his point is, it’s not really worth it, is it? But here’s his point. He said, at the moment, we’ve got engineers designing this, you know, they’re trying to optimise for everything for speed and all of that. What if you ask Disney to design a train ride? And I was like, that is such a good, it’s back to where we started this conversation on thinking differently, you know, trying to kind of think transformationally rather than incrementally. And it’s like, yeah, it was just a wonderful idea of like, you know, you’d have this train with ball pits and playgrounds and cinemas and whatever else going on. And yeah, and you obviously wouldn’t care how long you were on it for, because the time would fly by.
Neil Watkins
Exactly.
Yeah.
Kieron White
So yeah, that’s a good thought to keep in mind as you sort of, I think as you’re trying to reimagine business opportunities with AI, how would
Neil Watkins
That’s fascinating because I think just what was flashing through my head as you were talking about that, I’d heard him talk about the, you know, the HS2 thing, you know, save yourself, save yourself 100 billion just by getting the world’s supermodels to walk up and down, giving everyone a heat shut up for truth.
Kieron White
Yeah, yeah.
Neil Watkins
You just give them better Wi-Fi and drink and the people will be happy to stay on the train for an extra 20. In fact, they want the train to slow down, not speed up, which is the same thing with the Disney thing, isn’t it? But what was going through my head as you said that was, what if we got Disney to work on the AI adoption problem? You know,
Kieron White
Yeah, exactly, yeah.
Exactly.
Yeah.
Neil Watkins
If, you know, if we’re worried about AI adoption from Tech UK, from a government perspective, from, you know, how do we increase efficiency and effect, how do we get people to use the thing more? We need to be thinking more Disney than we do AI perhaps. So, there’s a thought.
Kieron White
Hey there.
Well, that would be storytelling, wouldn’t it? There is a whole bunch of the world of storytelling with technology. Even Gartner’s Data and Analytics Summit has sessions on telling data stories. So yeah, I guess there’s one angle on it. But yeah, well, there you go. Those are thoughts to ponder as you sip your beer.
Neil Watkins
Yeah, it does frequently.
Mhm.
Kieron White
Later on.
Neil Watkins
Right, well, for reasons that we will not explain to you, I will see you in another bar tonight and maybe we can talk about the Disney location of Leading AI. Well, there’s a couple of clowns.
Kieron White
It.
Perfect. Love it. All right. Yeah, bring your clown outfit.
Neil Watkins
Alright, fine. I’ll see you later.
Kieron White
See you later, bye bye.
Neil Watkins
Seven.
Neil Watkins stopped transcription
