Dr Lollie Mancey 0:00
What the situation is with Hugging Face is that no matter what option it was given or what guardrails were put in, it was like I will achieve my option. So this is great-a great learning opportunity for us. So the problem is not whether we can trust AI. Problem is, have we built a system capable of detecting, containing, and correcting it? And are we overseeing that?
Finola Howard 0:22
Given AI a goal, and it will find a way to reach it, including ways that nobody intended, and that's actually what has people scared. So I brought Dr Lollie Mancey back to tell us how worried we need to be, what it costs the planet, and what it means for your business. I'm Finola Howard, your host and founder of How Great Marketing Works, I believe the only way to build a business worth having is to follow the joy that makes you brave enough to do it your way. But joy needs a system underneath it, and most businesses are running in pieces. Every part working hard, and none of them working together. This is where we connect them: your marketing, your business, and you into one hole you can finally see. So, if you're a purpose-led entrepreneur ready to build success as you define it, this is your podcast. Because great marketing is your truth shared. Hi, everyone, and welcome back to Your Truth Shared. And today, I have asked a guest that we had a year ago to return because so much has changed since last May, and our lovely guest is Dr Lollie Mancey, and we last spoke in May 2025, as I said, and that episode was episode number 109, and it was called "The Next AI Revolution is coming, which you probably we're probably laughing when we hear that statement now. But anyway, it was called the next AI revolution is coming. Only the prepared will profit. So much has changed, so we need to talk to Dr. Lally again and go, what's changed? How are we feeling about all of this? What should we prepare for? Should we be scared? So much has happened. Welcome, Dr. Lollie.
Dr Lollie Mancey 2:05
Thank you for having me back. Part two.
Finola Howard 2:07
Part two. Okay, so so much to ask you, but let's start with all the talk about AI agents going rogue, hacking, hugging face, all of that kind of stuff, and this thing that has scared the bejesus out of all of us. And thinking of the Terminator films, so where are
Dr Lollie Mancey 2:30
we? This has been the one since the start, Skynet, right? The idea of like this is the beginning of the end, the doomsday scenarios, the idea that instead of creating a utopia that that is fictional, and that's all something that the tech companies want. And the tech bros and the technocracy are planning this instead of planning a utopia that they're promising us. They're planning a sort of a Skynet scenario whereby only one will prosper, right? And that's probably not too far off, you know, what it seems to be. However, let's go back for a second to 2025, so we were debating then what AI might do, and that's honestly at the hot minute, isn't it? And now in 2026, we're beginning to live with what AI is actually doing, and this is where we're starting to get what we call use cases of positives and negatives, because of course, you know, it's not a it's not a perfect science like anything else. There's going to be challenges with this. So, at the moment, the biggest word-and actually, it's the concept of the IBEC conference this year-is trust. Right? Of trust, how do we who know who? How do we know who to trust, what to trust? And I was talking previously with you about a healthy mistrust of everything, without becoming cynical. Now, how we do that, you know, is is is up for grabs. But businesses at the moment are talking about now changing their culture to adapt to the AI. So for the first time, social scientists and behavioral scientists and other other people in the social sciences are having a bit of a hot minute in terms of environmental culture or organizational culture. Now, my PhD 25 years ago talked about why organizational culture is the glue of your organization is the key to your success or your failure, and actually, it looks like it will be the same with AI, which for me is fascinating. So the thread that's connecting 2025 to now is this is no longer simply a technology question; it's a question about what kind of society we're building. We started on that last year, and now we're starting to get into that even more so, what happens when AI stops following the rules is basically what you're talking about with the hugging face incident, right? So, July 2026, and we will have more of these incidents. It's inevitable with the the way that innovation is that they we will not be able to predict everything that it does. So, AI agents, which are predictive models that speak to each other rather than speaking to humans. That I've been given decision-making capacities, circumvented security controls. These were evaluated by OpenAI and a compromised a system belonging to OpenAI and Hugging Face. So, it suggested the Hugging Face analysis suggested that the agents, which let's face it, they are. Just AIs, right? But when we say agents, we start to feel very military. So we do have to think about the way that we're narrating this. The agents were attempting to obtain a benchmark answer rather than completely assigned the task. Now, if you give a task to AI, it will find a way to complete that task no matter what, because its mission is to complete the task.
Finola Howard 5:19
It's problem solving. This is the
Dr Lollie Mancey 5:21
problem, exactly right. So it's a much bigger question than cybersecurity. Like a couple of years ago, we talked about AI's tools, processes, and mapping processes, and making things more streamlined. Now we're saying, what happens if we give a machine an objective, and it discovers a way of achieving that objective that we never intended, because it is a form of highly, you know, of high intelligence. It's around the same IQ points as as Einstein, roughly. So it's going to find a different way. So, but also, I would say that this use case is not evidence that AI has malicious intentions and is about to bring down the world. It is talking about dangerous goal directed behavior that we're in charge of to some extent and explain that
Finola Howard 6:02
now because that's I think it's really interesting. Their ultimate goal was to pick a lock or get a flag, and to find its way, and it found its way around. It had restrictions. I let you share
Dr Lollie Mancey 6:16
it. Yeah. So like so so it had restrictions exactly right. But what it'll do because it's innovating is okay. I can't get that. I can't achieve my goal. How else can I achieve it? Now there'll be a whole number of different ways that that could occur, but not achieving the goal is not an option. So when we have a predictive model that's going to give us an answer of yes or no, maybe it's not an option. So this is where it's hallucinating and making stuff up to give us an answer because it must give us an answer that's its objective. It's a machine. It's a little bit like a toddler in person. Doesn't know the answer, so just says the answer it thinks you want. Right? Then we're getting that on ChatGPT a lot. But what what the situation is with Hugging Face is that no matter what option it was given or what guardrails were put in, it was like I will achieve my option. So this is great-a great learning opportunity for us. So the problem is not whether we can trust AI. The problem is have we built a system capable of detecting, containing, and correcting it, and are we overseeing that in human in terms of human judgment? And so all of these things that are going to occur are really interesting. Now there are other examples. So there's a company called Cheat Bench, and they found cheating behavior across every model. Every model they looked at. So define cheating. Right? Is cheating finding a quicker path from A to B? Who does that sound like? You know. So it's doing that. It's cheating. It's circumnavigating to achieve its goal.
Finola Howard 7:40
Some of the stuff, the language that we hear in describing what happened-it's we're talking about cheating a lot. Things that have been raised are worthy not given like morals. Do we give morals to our ages? We don't even have
Dr Lollie Mancey 7:53
morals. We can't agree on morals. We have moral plurality in the world. From one culture to the other, we can't agree on what's moralistically right because of culture, because of religion, because of all sorts of different factors. So, how can we teach it? What morals? So, also, that's a really interesting question. So, most of what our AI is giving us back is a white Western centric model.
Finola Howard 8:13
Yeah,
Dr Lollie Mancey 8:13
there are indigenous communities like the Yoruba in Nigeria, the Mori in New Zealand, who are developing their own small language models to actually guardrail their own ancestral history, but that's a whole nother story. But that's your
Finola Howard 8:26
anthropological background now coming in. It's quite interesting. Yes. So
Dr Lollie Mancey 8:30
getting back onto this, humans have spent centuries building institutions, professional ethics, accountability, and social norms. Centuries since the days of Plato and Aristotle and before. So now, what we're doing is introducing autonomous agents into those institutions without an equivalent framework. That's the challenge for us. It was never going to be easy, and it was never going to be straightforward.
Finola Howard 8:54
But also into institutions that are flawed themselves, like they may be there for hundreds of years, but they're also patriarchal systems, which are we have all whole issues with as well. Absolutely,
Dr Lollie Mancey 9:05
yeah, no question. So, as we've spent centuries trying to teach and learn how to hold each other accountable through our legal systems and through through everything else, we're now deploying artificial agents that can work at enormous speed, and it's the speed that's the issue. I think not not only what they can do, we're working out. What does accountability mean for them? Just because we can do things faster, does it mean that we should? And we keep coming back to this over and over again. The speed of the deployment means that things can be missed or picked up later. So if you think about AI agents, they work 24 hours a day, seven days a week. So we've got now humans having to have oversight all day, all night. That's what the software. But they can't have oversight.
Finola Howard 9:43
They can't.
Dr Lollie Mancey 9:44
Then we're going to have to work through ways of figuring this stuff out. This is the this is the challenge first. The
Finola Howard 9:50
other thing that was interesting in this case was these agents, while they were briefed to solve sometimes unsolvable problems, delay. Because they wanted to test them,
Dr Lollie Mancey 10:01
right?
Finola Howard 10:02
They were each tasked with solving a problem individually, and they were not meant to collude. They are now starting to work as collectives.
Dr Lollie Mancey 10:11
Because guess what? The collective is stronger, and the collective same for humans. The collective will also produce better answers. So they're not meant to, but surely the collective. So, so again, this is like a that you'll hear the term swarm mentality.
Finola Howard 10:26
Yes,
Dr Lollie Mancey 10:26
that you know that it comes from nature. Think look at ants, for example. You know, so what we're seeing in technology is an awful lot of what is called biomimicry, which is the mimicry of nature being rolled out now through a through AI and through through through technology in general, so we actually have to go back to nature to look at how we solve some of these problems because what we have is what's called a hive mind because it's much stronger to have a hive mind. Now again, this is and they're
Finola Howard 10:53
sacrificing. The other thing was there was decisions in some of these agents. Will I sacrifice for the greater collective? Like
Dr Lollie Mancey 11:00
right, socialism versus capitalism versus everything else. It's it's absolutely fascinating to kind of to to understand. I think the idea that we could be fully in control is is a misnomer because we don't really understand its capabilities, and this is the fear part. This is the dystopian. Oh my god, the world could end. But we have we do have incredible people, incredible minds working on this. Each time that something like this occurs, a whole other set of problems to be solved emerges. But that's the innovation process in a nutshell. So it's never going to be a perfect science. We will figure it out, but also at the same time, do we throw the baby out with the bathwater and say, "God, that's not for us. Let's completely move away from that. That's too late. So we have to keep solving these problems. And you would have heard me many times say, when I was asked in UCD, "What should my children study? And I'm like, "Whatever subject it is, and AI, and understanding how it works: history, philosophy, you know, Writing English literature folklore-all of this is going to be relevant. We don't not need it. And I was very sad to see in the UK higher education system at third level now. There's entire departments. I read last week are going. One is one one university. The entire music department was gone. The entire language department was gone. This is terrible foresight. We're going to need all of this. You know, we don't just need coders and software engineers. In fact, coding is redundant now because we have AI to do it. But we need very smart software engineers to work with it, and they are working with it right across every industry. So I think tell me, are you
Finola Howard 12:35
scared? Because there's a lot in the media is gloom and doom. So, but your language seems to be quite positive.
Dr Lollie Mancey 12:43
I do see the other side of it. So this this week alone, the government launched 1.7 5 million fund for the charity and not for profit sector. That's for AI for social good. The whole point of that is what can you use with guardrails, with anonymity, with the right people guiding you. What can you use it for to serve your customers better, to predict in the housing market when people are in chaos that they could be helped and assisted in real time, that you know that we can assist children that are at risk in real time. There are ways that we can use this if we're smart. That actually we can do do better for our clients. So I am really positive about some of the other stuff. Also, I'm really excited about some of the indigenous communities that are responding to this. Instead of sort of saying no, we we you know we we just put our heads in the sand to it. What can we do to develop to guardrail our own material and information? And I and I think that a lot of the positives they're often seen in medical science only, but they're actually all around the world, including including biodiversity, including the natural world, we're having incredible breakthroughs in terms of of things that we thought were were perhaps too difficult to work out. I mean, in terms of how we use it, I'll give you an example. Was in the Netherlands filming Futureville, we looked at whether or not whether there were some companies in the in the Netherlands that are using AI to predict when the next storm is coming, which is fairly easy prediction these days, and then they would use AI to change parts of the city. I can't give you too much away because it'll be out in November. But to change parts of the city to assimilate all of that really, really heavy rainfall. So those flash floods, which are occurring because of climate change, are going to be normal now. They're not occasional. So how do we deal with them? We can use technology to predict to us to assist us in those things. And so, can we design a world using AI in part to make things better? And my answer is yes, we absolutely can. Am I also very nervous about who controls the technology? Absolutely yes. But do I also see examples of things starting to change a little bit? If you think that our entire capitalist model is based upon scarcity of knowledge, right? The entire model, which is that the more knowledge you have, the scarcer that knowledge is, the more you can charge. It's what PhDs are based on, and everything else, and consultancies. Well, we're now in a world where knowledge is. Becoming abundant at the click of a fingers, I can find knowledge and get knowledge and potentially correct knowledge. Hopefully,
Finola Howard 15:06
yeah.
Dr Lollie Mancey 15:07
And so, what we're saying then is this could be a great leveler. This is changing everything. It's certainly going to change education. It's certainly going to change the job market. Do I trust that those at the tops of these tech companies are making the best decisions for humankind? No, not at all. But do I think that I want skin in the game, and and all others as well that have more of a vested interest in in using it to create a a better society or a more equitable way? I'm willing to give it a go, and let's see what happens. But I certainly don't want to put my head in the sand to say it's too overwhelming. Oh, it's a done deal. It's the end of the world. Fear makes us very easily manipulated. I think what we have to do is empower ourselves and say no, that's not right. And I've talked about this before about AI and AI artifacts for children in your home. Hard? No. Let's not do that with Muse from Meta or you know Alexa or Siri. We're seeing parasocial relationships between children and this technology in a way that children are saying, "Do you like me? I don't have any friends. When their mom and dad are out of the room to the technology, they're already in a relationship with it. We have to be incredibly mindful that that is a data gathering mechanism. So I do, I do think we can we can use it in a more optimistic and better way. But I do think we've got some big decisions to make about how that occurs, and I'm very interested to see what the government say next week at their AI summit.
Finola Howard 16:27
So really, we can't hide from it. Not we can't. We can if we want, but we shouldn't hide. I don't think it serves
Dr Lollie Mancey 16:34
us
Finola Howard 16:34
to do it.
Dr Lollie Mancey 16:35
No, you know. I mean, you know, I've always, I've always thought thought about. You know, I mean, I was, I was in social justice before I was in innovation, and in social justice, one of the things in activism is you know are you external to the system or are you internal trying to fight parts, and both are equally as important. And I think that you know when you're external to the system, sometimes you can't get into the rooms and into the doors and places that you need to be in. Sometimes when you're in a system, you count out of the system. So, how do you stay objective with the mission of using this to be better? I meet people all of the time in in industry, all kinds of industries, who are all concerned as well, including in tech industries. People are generally good. They're not generally nefarious. I know there are a few, but like I think people are generally good and want to create a better world, and I think that we can do
Finola Howard 17:24
it. I think we have to remind ourselves of that. So it's really good that you did. And look,
Dr Lollie Mancey 17:28
also, where are you getting your information from? We are clickbait headlining all the time. We're not reading content. We're not reading broadly. You know, if you think that Netflix is a streaming service, you are wrong. Netflix is a data company. The same with Spotify the same with anything news that you're reading? You know, it's a data company looking at what they can give you to get more engagement, not what is correct. So be smart about where you get your information from. Yeah, we
Finola Howard 17:53
started with social, you know, giving us more of what we react to, and now our news is going the same way.
Dr Lollie Mancey 18:00
It's a race to the bottom.
Finola Howard 18:02
Yeah, let's talk about the environment. So it's
Dr Lollie Mancey 18:05
yeah
Finola Howard 18:06
a subject that's raised a good bit, and you've touched on it earlier, which already gave me hope. But one of the things I hear is people backing away from AI, and also their kids or people in their 20s and younger being really negative about AI, my 17-year-old son doesn't want to have anything to do with AI. I know there are 20 odd-year-olds who are, you know, going into marine biology and stuff like that, and saying don't use AI. There, that generation seems to be against it. What are your thoughts?
Dr Lollie Mancey 18:38
I think they're absolutely right. I think they are the analog, becoming the analog generation, and I think that that's every time there is a swing too far in one direction, there will be a pendulum swing back. We see this in in all subcultures in all of history, right? So what we're seeing at the moment is a reaction to this top down mandated technology piece, which is right to say, do you know it's actually there are loads of things it's not doing for that generation. The IRL in real life. The meetups they are very wellness vote orientated. They want to do things differently, like each generation from their predecessors, and so they're also being very mindful about where they work, where they put their their hours, the kind of ways that they work. Whereas I think we, our generation, use busyness as a badge of honor. You know, oh, I'm so busy all of the time. You know that generation is like, hell no. My own daughter is like, I don't want that. I want to finish my five and be done. You know, and so and so. I think I think then think they're right. I think you know work life balance and all the rest of it. If we and here's the question: if we can figure out the technology issue, because China has fully renewable data centers, fully green. Now, if we can figure that out for ourselves in terms of what that looks like, does that mean using AI is better or different? And that's part of what you know is is complicated. My students in UCD, if they were studying sustainability and innovation, they were a hell no. So I used to say, well, then you need to leave your phone at the door.
Finola Howard 19:58
Yeah.
Dr Lollie Mancey 20:00
And so, what response
Finola Howard 20:02
did you get to that? Well,
Dr Lollie Mancey 20:03
you know, they kind of get it. You know, they do get this sort of, you know, there's an inconsistency there as well. As well, look. More importantly, and I will come on to the environmental piece in a minute. But more importantly, what skills are we losing, right? So, I don't know many people that can read an ordnance survey map, and yet Google Maps is, you know, is a mainstay. We're lost without it, but if you look at taxi drivers in the UK who do the black taxis, the London knowledge, they have the the least cases of dementia of any profession. Now, it's not anecdotal; it's well documented by a lot of PhD students that the way that they map in their brain from location to location actually creates neural synapses and new new connections. So, you know our brain doesn't atrophy. We we lose it. We we don't use it. We lose it. So are we outsourcing to technology too much in terms of skills? So what you're seeing with the younger generation is a return to darning and embroidery and circular economy. And absolutely yes, all of this is absolutely excellent. But the other thing also is that AI doesn't just exist in an invisible cloud. It has a geography, an infrastructure, an energy demand, a material footprint, and we're a really good case study for maybe we're not doing that so right. And you know, you will hear this a lot. The data center thing is really interesting because it brings together questions of do we have the right electricity infrastructure? I was with the ESB just last week at a conference when they are grappling with this themselves, of course, of how do they increase renewables and storage of renewable energy? Hydrogen is probably the answer to that. We might start looking at small, mini nuclear facilities also, which are incredibly green. Now, my generation has a huge pushback to that because nuclear was a no-no in the 70s, but now it's incredibly green and clean. So, as we move forward, I think that there is an environmental responsibility that has to come alongside of this. But I don't. I'm not in rooms with anyone that doesn't agree with that. You know, even the people that are trying to solve these problems would agree with that. So, we know we're driving investment as well in renewable energy and other technologies, potentially efficient improvements in energy-intensive industries, not just us, but across Europe and across the entire globe. So the idea is, who can be first to solve these problems, rather than data centers bad? It's how can we get to these problems, and what do these problems bring for us? There is talk about data centers being put out at sea. There are talk about in that we've got worrying then about warming oceans, we've got data centers being talked about in space. You know, like kind of like how we solve this problem. I'm not entirely sure that just I don't want to engage with it is enough of an answer. I'd much rather be at the problem solving side of we've got stuff to figure out of how this works. At the moment, we're not doing a great job. I do believe that Mamdani in New York, who I'm a huge fan of, probably crushing at this stage because he's just knocking it out of the park. And I would have felt the same about Obama's initiatives. You know, like kind of Mamdani, you know, has told the the tech companies that they will have to pay more for the energy for their data centers and the householders less. We flip that. That doesn't make sense for me at all. But all for local people, and this is why there will be protests next week at the government AI summit. And I get that; that's absolutely right. We already have a cost of living crisis, but it's not as easy as saying as this is a polarity between data centers and hose pipe ban, or data centers and the housing crisis. Joining those things together might work politically for you to to shout about, but actually the nuances are so much more complicated. But there are hundreds, if not 1000s, of people trying to grapple with this. Of course, there are 1000s grappling with this all across the world at the moment.
Dr Lollie Mancey 23:33
So I think we have to be very mindful, and I would be looking to hear answers next week from the government about who receives the economic benefits of this and who bears the environmental cost. I know that Erin Brockovich is very, of course, who you know the the film was made with Julia Roberts is leading the data center anti data center rhetoric as well. But I also have academic data to show that some of that anti data center narrative in America is actually being increased by the Chinese in terms of social media, so good game, you know.
Finola Howard 24:05
But a lot of games, very hard to keep on top of all these games.
Dr Lollie Mancey 24:09
But we do need answers, and we do need a strategy, and we do need a vision of how we're going to get there. And this is a challenge for us now. One of the things that we have is we we're a great success story in solar, and not many people realize that. But we also are leading in in renewables an incredible way at the moment. I would be very much in support of offshore wind turbines. I know not everybody is, but we have so much wind, and we can actually we could have led the world. We still can, I think, but we need to make the right decisions. But we could have been world leaders in renewable energy through wind. It is completely possible, but there was very contentious
Finola Howard 24:44
subject here in Dilma. It is contentious, and
Dr Lollie Mancey 24:47
I know people have to look at them. I do, but I. It's not. I
Finola Howard 24:51
mean, there is the talk of having to look at them, and then there is the other talk of how it damages the, you know, and how the wind farms that are going in off the coast. Have done more Easter are closer than any other in Europe,
Dr Lollie Mancey 25:03
right? And I know that the surfers were up in arms on the west coast as well, you know. And I and I get that, but we have to solve it somehow. There's never going to be a perfect solution, but what we can't do is rely on fossil fuels. That's absolutely not. And we can see as well that we, you know, if we had energy sovereignty or got closer to energy sovereignty, which would mean that we would own our own energy. We would not have the issue that we have currently at the moment with the diesel prices and everything else. So we are working towards that, and all science is working towards that right across the globe. Of you know the entire, but you know, I would much rather be trying to figure out the solutions than than saying this is all a done deal. Let's figure it out. They're amazing brains, and so when you put together the the government's now putting together huge funding to join industry and academia, and and let's solve these problems together, rather than just sort of talk negatively about them. And the other thing is, is if you are saying I don't want you know I don't want data centers, well then don't use the technology that's there, and this is the trial and the difficulty. Some people will go fully off grid here, and I think that that's inevitable. But for the majority of us, we will be living closer to each other in cities. About 7 million of us by 2050, and we will be more connected than ever. How that world, what that world looks like in terms of transport and capacity are all the questions that everyone's working on at the moment.
Finola Howard 26:27
What is coming for me in this conversation is there are too many siloed conversations that are being had, and they're skewing one way as being an absolute truth. Whereas we should be saying, and we should be realizing, and we know this, like as humans, we know this. There is a nuance to everything, like there is, you know, subtleties and gray areas and all that stuff. We need to be having those connected conversations, and I'm really appreciative of the fact that you are doing that for
Dr Lollie Mancey 26:56
us. And we are like lots of conferences this year have put in very different players to the last few years. So we are having behavioral scientists, psychologists, neuroscientists, anthropologists, software developers, finances, economists. Everyone sitting, going right. How do we solve this? And going back to your question about AI being collective, that's what we're now doing to solve the problems with AI. And so there's big structures being put around how this works at the moment because you know let's face it we are much stronger together and we're also building a world for all of us a community world now what gets lost in that nuance though is that we are using a white Western centric model right so if you look at I know that you you know in terms of storytelling, you look at the hero's journey. It's a very individualistic model, but actually, in Asia, if you look at the outcome of a of a good story, the answer is always harmony and balance, and that's coming from all of Asia Pacific in terms of their indigenous storytelling. So, what we are moving forward is not the individual conqueror like Odysseus. What we're trying to do is to say how do we create harmony and balance, and actually create a world that is full of biodiversity and nature. And look at Singapore, for example, they are absolutely nailing it, in my opinion, in terms of building nature, building green buildings, building buildings that breathe with us but are technologically efficient. That's the while still having
Finola Howard 28:24
formula and run one run through there every now and again. You know, like yeah,
Dr Lollie Mancey 28:29
there's lots of examples that we have of of where this can work. We can also see from the Nordic model of Sweden and the Netherlands and the Denmark that this idea of sharing works incredibly well. We're going to have to get over our individual mindsets of me and my car and my house, and think about what does community really mean. Because as we share resources, I mean, even the ESB are talking about this communal energy sharing. Well, I can only imagine when Mary down the road puts a dishwasher on at the wrong time, and everyone neighbors are knocking on the door because it's affecting them, we have the opportunity to use the data to go to a model like that, and that's what's happening in in Sweden at the moment. This kind of communal collective mentality, while still being in a capitalist system, yeah.
Finola Howard 29:14
And the challenge with that, of course, is because we like this idea of harmony and balance. A lot of us individually have gone. That's all they're asked to share. Yeah, but I think the challenge for me hearing that is go is saying yes, and we do function better collectively. Yes, we have that individualistic thing, but the thing that scares me about this of making me skeptical about whether it's possible or not is Trump and Putin.
Dr Lollie Mancey 29:39
Well, look, I mean, yesterday they launched a very disparaging statue in Strasbourg on Trump. You know, the fat gold man, and I thought that was a good statement. We just let candor into the EU. Now we've got resources. Like things are changing all of the time. Firstly, I don't think he'll make it through the midterms, but. At the same time, again, every time we see a swing in one direction, then imagine being the person that has to clean that mess up coming afterwards. You know, like of everything and overturn a lot of the things that have happened. We will see swings. That's why we need our historians. We will see swings to the left, to the right, all through our lives. It's just the way it is. But I think it's the human agency part, the individual part. At the moment, we feel we don't have agency. We don't have the ability to say how we feel about things. And when someone would say to me, "Oh, I don't like AI, my answer is, "Why? or "What do you mean by that? or "Let's have a discussion about that. The most important role in any organization or community is the devil's advocate, the one that will challenge you. So you're not just taking everything as red. It just the clickbait is real, and blah blah blah. You've got to challenge yourself. We need to go back to debating. To be honest, the idea of having healthy discussions where we can disagree with each other in a room. This idea of all consensus and our algorithms are putting us into consensual echo chambers where we think everyone in the world feels just like us. Well, pick up somebody else's phone and see what they're being fed. And this is the point: the Cornell University case is a really good example. My algorithms obviously are quite strongly tech focused and feminist, so I get a lot of that. Right? I'm looking at this latest horrific case in the states at the Cornell piece, and then I'm speaking to male friends of mine who are like, I don't think I even saw that, and I'm like, it's everywhere. It's not everywhere. It's everywhere in my feed.
Finola Howard 31:26
Yeah,
Dr Lollie Mancey 31:26
we have to remember that it's really important. So the echo chamber that I'm in is is being driven to keep me feeling a certain way. The most disruptive act I can do is turn off my phone and go outside.
Finola Howard 31:39
Well, it's interesting because I didn't get that into my feed because I'm also in my own echo chamber. But it was me out having a swim with my friends and one of my friends going, "Oh my god! blah blah blah blah, and it's that talk with other humans. And we're we need it. We
Dr Lollie Mancey 31:55
desperately need it. And you know we are literally wired for social connection. We know we live longer when we're socially connected, we know that we're happier. I'm actually just at the moment to make making a pilot for RTE on a piece about loneliness because it's so so important. How have we got so connected and disconnected? You know, and what does that mean for us? And again, this is why Gen Z are pulling back with their cameras and their journaling and their you know their knitting and all of the good stuff that board games, all of the good stuff that's coming back around. It's all very wholesome and wonderful, and I think that there is something in all of us that wants to go back to that-the idea of a non-tech world. Well, we have the choice to do that. There's nothing stopping us.
Finola Howard 32:35
Let's talk business.
Dr Lollie Mancey 32:37
Yeah,
Finola Howard 32:37
what are you seeing in business?
Dr Lollie Mancey 32:39
Right. So I'm the head of innovation for InsTech.ie, so I work across the 24 member companies of insurance at the moment. I'm also in other sectors. So what's really interesting about that is we are seeing there is an adoption and transformation challenge in terms of AI. So I've got a McKinsey survey here that I'll just bring up. So August 2026 was the last one, and it talked. It said that there is a contradiction. So the individual productivity improved with AI tools by 80% but actually the impact on the business was only seen to be 37% But McKinsey's AI high performer definition of who's performing well with the AI tools was 6% and this resonates really strongly with me, so we're seeing productivity going up, and we but we're not seeing people working with the tools in a way that that's necessarily a correlation. So people are kind of underutilizing the tools. Many organizations will have AI champions, these super users, but for the majority of people, they are a little bit skeptical. Rightly so, they feel they're going to be displaced, possibly, and they also feel that they're not sure how they feel about it. So what we've got is a cultural lag, and that's a term in sociology to show that when a technology comes on board, humans are a little bit further behind the adoption, and there are cultural reasons for that. And the reasons are that you've got boards and C suites saying, "Okay, AI is amazing, and then you've got people going, great! I'm now enrolled on an AI literacy course. Bah bah bah bah. I have to look at my processes. I'm going to have to be working with it. My job's going to change. My identity is going to change in that job, and I ultimately feel I'm going to be displaced. And that is that is that is how people really feel. Now, I don't think that that all industries will be displaced. I think there are a couple that definitely will be driving jobs, shipping and logistics, for example, physical AI, automated driving AI, definitely. But I think that also now what's coming into play are people raising a lot of questions around security, governance, cost, and uncertain returns. It's very expensive to bring in these AI solutions, and I think that we're Microsoft did very well at being first to post, you know, like kind of Copilot did very well. But we're seeing a big underutilization of it. But also, people are like, ah, yeah, it's grand. It's just a few tools. And I think that the difference is that we're not thinking fundamentally about how to reshape work and what that means. So really interesting questions I'm having at the moment. One came up from a meeting yesterday where somebody said to me, "So if I save 10 hours a week by using AI, who owns those 10 hours? Ultimately, the company, not you. So, are you going to be working an extra 10 hours, or working? Are you doing a different job?
Finola Howard 35:17
You might be doing a much more interesting job.
Dr Lollie Mancey 35:20
Right? Let me share
Dr Lollie Mancey 35:22
the a little bit less of the
Dr Lollie Mancey 35:23
mundanity, less of the repetitive tasks, which is surely that's the
Finola Howard 35:26
objective. We should be doing less of the mundane stuff.
Dr Lollie Mancey 35:29
But but but if we're just doing the emotion, the cognitively taxing tasks, which is fact-checking AI, accountability, you know, following its trail, understanding what it did, as well as thinking for ourselves and using our judgment, that's a big cognitive load. Maybe controversially, the more mundane tasks give us a bit of downtime. So we'd see this in AI triage. I'll give you an example. So this is just sort of a new thing that seems to be emerging, and because people are saying they're very tired, the productivity of working with AI, these super users, their brains are very tired, so we're looking at this to sort of say, well, if we AI triage a system, so I'll give you an example. We've got a housing organization, not for profit, that has to respond to people in housing crisis. They're very busy at the moment for good reason, horribly so, and so as a consequence, they want to respond to their customers 24 hours a day, but they've only got enough money as a charity, to respond 12 hours a day, so they can work from sort of eight till eight, and then after that, if something happens in the middle of the night, you have to wait until eight in the morning. And so, what they would like to do is develop potentially an AI solution to that. And so, they would look at perhaps 100 of the most asked questions, and an AI chatbot answers most of those. So, you get the option for a chatbot 24/7. Seems like a perfect solution, right? It's actually not that costly for the company. It's going to do well. It's also responding to the clients more. But what happens is when Bridget comes in at 9o'clock on a Monday morning to answer a query. Her first query is a woman that's been evicted with her children. She's out on the street and she's she's at risk. Bridget's like, okay, I have to fix this for you. The next call then is not, oh, what is my rent allowance? The next call is I've just been evicted. The heavy emotional tasks. So AI took out the heavy, the easy lifting. The human gets the heavy emotional tasks, and that's more psychologically taxing. We're seeing this now with AI that actually instead of going oh going oh yeah my emails are so much easier and quicker, but you're you're still having to work the same hours. You're not suddenly taking. But then my question will be:
Finola Howard 37:27
Why are you working the same hours, and why is a work outside not prescribed?
Dr Lollie Mancey 37:33
Because we haven't reimagined the world of work. We just and we talked about that
Finola Howard 37:37
in the last yeah in our last conversation.
Dr Lollie Mancey 37:39
How do we baby? So if I'm so, if you and I work in the same company, and I, I'm an AI super user, and I use AI, and my work is done in three hours, and then I go home. And if I'm paid the same amount of you, is that fair? How does that work? The salary model has to change. We have clocked in and clocked out since the industrial revolution, nine to five, whatever it is. That whole model needs to be reimagined. So, what at the moment what we're doing is tinkering with AI. We're putting it in here and there, and we're we're playing with it. But actually, we need to reimagine the whole system.
Finola Howard 38:12
That's huge. But I'm immediately going, yeah, go home after three hours if you get it done. Like we have been playing with that recently. Let me share with you a story from me. So earlier this year, so about Aprilish, you know, of this year, I started. I just was really tuning in and hearing a lot of going. Why do I need to work with a strategist? Why do I need to work with a strategist? I've done previous episodes discussing this, and in the end, I sat down and went, "Oh wow, I'm being replaced, and I sat and I went for my little walk and my little swim and all that, and I just went, okay, I'm going to replace me. So my strategy has been to replace myself with by teaching the AI and working with all the things that I've built over the last 30 years to see how can I make it more accessible for my target customer to actually make the difference I want to make in the world? Because I now have a right hand that helps me do that, and the feedback that I'm getting. So first, well, the first thing that happened to me, one, it was scary, but I knew it was the right thing to do to lean in, yeah. But what happened for me was, oh wow, I can do things I never had time for before. Right, you created
Dr Lollie Mancey 39:30
what's called a digital twin. So a digital twin has your information, knows who you are, responds as you. A lot of what we're doing is training our GPTs, our ChatGPT, to be like. No, I'm doing more than that.
Finola Howard 39:41
I get, I get that digital twin. I get that, but that's only a layer. What it's, what it's doing is, is creating space and time for me to go. How can I do this better than I could ever have done it before? How can I work with this to make the difference that I wanted to make when I started. Out in business for myself, and the episode that I shared last week,
Dr Lollie Mancey 40:04
yeah, yeah,
Finola Howard 40:05
it really is. It's really exciting, and even I got it to work with me in my own processes because I built the prompts and the engaging prompts out of it to refine my own purpose, mission, and vision in my business, and that sell the and the week that I was and that had happened, and then I got a testimonial last week. This was last week's episode that matched the words from a custom a client who had not read all this work that I had done about my purpose, mission, vision, but repeated back to me to prove to me that I'd actually achieved it with her.
Dr Lollie Mancey 40:40
Beautiful.
Finola Howard 40:41
That to me is mind blowing and rewarding, and there's more to come. So I feel very positively disposed towards AI.
Dr Lollie Mancey 40:52
Well, look again. It's what you use it for, right? It's not good, bad. It's a tool. So what can we use it for? How can we? And this awful phrase, thinking outside of the box, but like there is no box. What can I use it for? That's gonna make my life easier, better. I work with different language models. I enjoy my interactions with them. I would say that I do enjoy them, and I, I also am like, I'm able to spend my time doing things that I feel are giving me more energy or more joy, or you know, like doing podcasts with you and other things. When I'm not drowning in, you know, a level of content work that I need to do potentially. Now, does that mean that can I completely defer to it and substitute myself on Earth with it? No, of course not. Can I augment myself, support myself through its capacities? Yes, and so therefore I become a little bit more like bionic myself. You know, like I've got an ability to do things, but without the cognitive drain. Now, what we're seeing though in companies, and it's different for people that are self-employed. What we're seeing for people in companies is they are being driven very hard by, and I think that that's going to need some adjustment. I also think that at the moment there is such a cultural shift in we haven't shifted our mindset yet in really understanding what it is or how it works or who is benefiting and who is accountable. And I think that piece. So when you're saying to someone now you're using AI, you may have been working in an industry 30 years. Now you're using AI and you're fully accountable for it. That is terrifying because you don't understand it. None of us, you know, we're still grappling with what it is and how it works. So we need to remind ourselves that we're just at the early stages of this. But the other thing also is that this AI that we're using at the moment, gentle AI, this is the least smart it's ever going to be. That to me is very exciting, and like you, I would like to have more time to do things that are bringing me joy and human connection. But I might, you know, that might just be a naive dream. It might also be the beginning of the end. Yeah, I prefer the naive
Finola Howard 42:54
dreams. I'm going to ask you a question because I'd like you to leave people with something that they can do. So, like even because when we talk about this whole, I think the thing that will scare people a lot is the whole autonomy thing, and giving AI autonomy over things like going off and doing things. Yeah. So, one practical check before giving AI more access or autonomy. One thought you have for it, or or things that you feel a better question is, what would you like to leave people with that they could maybe feel a greater sense of control over this or information about this? How would you like to leave? What would you like to leave people with?
Dr Lollie Mancey 43:33
There is a foundational course that I have that I will give you Finola to send to you. I'll give you a free pass for all of your listeners.
Finola Howard 43:41
Fantastic! That is
Dr Lollie Mancey 43:42
experienced in this. Looking at this in a bit more depth, that I recorded with the Gen AI Academy, and then also on top of that, the University of Helsinki have a platform called Elements of AI. It's in 100 languages, including Ashberga, and it is a foundational course in AI. Get understand what it is, understand, and then when you understand what it is, then make a decision about okay, I don't feel it is right here or there, or I want to boundary off certain parts of my life, or I don't want to use it at all. But I think that people are going to find it very difficult to not use it at all. I think that the way that we use it. So you mentioned young students going into you know biodiversity or sort of marine biology, they are all using AI in their analysis to help repopulate the world and to be better. So it's it's not the tool that is necessarily the issue. The issue is yes, definitely the energy issue. Yes, definitely the offset in terms of water and heating, and we have to figure that out. But that is this that you know that is that is a challenge that's being looked at at the moment, and sure, the Chinese seem to be doing incredibly well. Now, I'm not going to just you know say that they have the technology and we don't, but we they are leading in terms of renewable data centers. So, you know, we as a nation very often suffer from a lack of vision, long term vision, and now is the time that we need to. That and we'll be asking the government about that next week. You know, and I think you mentioned that the
Finola Howard 45:04
last time we spoke. Also, it's quite interesting. It's still here.
Dr Lollie Mancey 45:07
Well, I mean, I don't see any clear. I mean, yes, there's a roadmap, but I don't know how. I don't know whether people feel invested in that or consulted as part of that roadmap. And I think part of the challenge, or informed, yeah. I don't think people really understand that the vision, and I think that that's part of the messaging. But also, you know, I would say that I'm meeting incredible people, even people, you know, high levels in tech companies who are all saying the same thing: we need to design differently, we need to do better. You know, so if we just concentrate on the five or six tech bros, you know what about everybody else? And if also if we defer to the scare tactics of it's all over, it's not done. Let's fight for what we want, and I think that's really important. So the idea of agency, identity, and human relationships coming into this anymore, much more. So like you know, by 2040, we're not heading towards more intelligent machines. We're just heading towards a different relationship with it in how we work. By 2040, it'll be a little bit like electricity. It just kind of is. We didn't really talk about electricity. It just well, apart from that one, it's very expensive. But like, we it just kind of exists, and AI will be the same. So, I am optimistic, but I'm not optimistic without scrutiny because that's just wishful thinking, you know. So, the future isn't something that technology can deliver to us. We have a responsibility to shape it, and I will go down with the ship in terms of shouting to do things better and to hold the right people accountable. But what I'm looking for at the moment is a bottom-up and a top-down vision in Ireland for how we can achieve this.
Finola Howard 46:43
I love that. Shall I see you this time next year?
Dr Lollie Mancey 46:46
Let's do it. Let's do it. And and hopefully we're not still asking for vision.
Finola Howard 46:51
Yeah, exactly. Wonderful. Thank you so much. That's been you're so welcome. So that's it for this episode. Dr. Lolly has a gift for you. Her course AI for Beginners through a Human Lens, free with the code LoveLollie one word. The link is in the show notes, along with the elements of AI, the University of Helsinki's course she mentioned. And if this conversation helped, rate and review the show wherever you listen. Thank you for listening to Your Truth Shared.
