Ep90 Leading with Humanity in the Age of AI w. Simon Defoe of Vodafone

AI is rapidly reshaping how organisations operate, make decisions, and develop talent. But as technology accelerates, what happens to the human skills that drive performance?

In this episode of The Chief Psychology Officer Podcast, Dr Amanda Potter and Angela Malik are joined by Simon Defoe CPsychol, Learning and Performance Senior Manager at Vodafone, to explore the intersection of AI, psychological safety, and leadership.

Drawing on expertise in organisational psychology, resilience, and decision making, the conversation moves beyond technical hype to focus on what matters most for HR and business leaders. 

How do we integrate AI into talent strategy without eroding accountability, critical thinking, or human judgement? 

What are the risks of bias amplification? 

And how can leaders protect belonging, trust, and performance in an increasingly automated world?

The discussion examines AI in HR, human centred leadership, board level decision dynamics, organisational resilience, and the role of psychological safety in enabling innovation without losing ownership.

For leaders navigating digital transformation, workforce change, and AI enabled performance, this episode offers a grounded and practical lens on keeping organisations both technologically advanced and deeply human.

Episodes are available here https://www.thecpo.co.uk/

To follow Zircon on LinkedIn and to be first to hear about podcasts, publications and news, please like and follow us: https://www.linkedin.com/company/betalent-by-zircon/

To access the research white papers mentioned in this and other podcasts, please go to: https://www.betalent.com/research
 
For more information about the BeTalent suite of tools and platform please contact: Hello@BeTalent.com

Transcript

Setting The Stage: AI Vs Humanity

Angela Malik 0:04

Welcome to the Chief Psychology Officer Podcast, the show where we dive deep into the psychology behind leadership, business, and success. I'm Angela Malik, and today we're joined by our very own Chief Psychology Officer, Dr. Amanda Potter, and our guest, Simon Defoe, Senior Talent and Assessment Manager from Vodafone. Today we will be debating the balance between AI and humanity from a talent and HR perspective. But before we jump in, make sure you hit subscribe so you never miss an episode. And if you like what you hear, please give us a rating so that more people can discover our trove of insight. If you want to keep the conversation going, connect with us on LinkedIn. Just search for Angela Malik or Dr. Amanda Potter. And as always, you can visit the CPO.co.uk or BeTalent.com for more resources on everything we talk about here on the podcast. So let's start with you, Simon.

Simon’s Role And Talent Lens

Angela Malik 0:52

We know you well from our work with Vodafone as a client, but for our listeners, could you introduce yourself and your role?

Simon Defoe 0:58

Hi, yes, so nice to speak to you both Angela and Amanda. I'm Simon. I look after all things talent assessment related at Vodafone, and that can be anything from what we do for entry-level roles, for graduates, for frontline customer care roles, all the way up to what we do for our senior leadership selection and development. And very much see my role as thinking about how we support decision making for individuals, whether that's for their own development, whether it's for managers making decisions about selection or promotions, and also for the organization kind of shaping that talent agenda.

Angela Malik 1:34

Amanda, with AI seemingly becoming increasingly human as the technology gets better and better, I can't help but think that this topic is very timely and lots of listeners will be interested to tune in. What interests you about the topic of AI and humanity?

AI’s Promise: Speed And Creativity

Dr Amanda Potter 1:49

This is not a topic I would have expected myself to record. So I'm a new adopter of AI. Well, I say relatively new. Over the last couple of years, we've really started to embrace it and use it within our organization, and it's absolutely pivotal to our strategy. And I personally find that I learn a huge amount from using AI. And it was from our conversations that we have with Simon that we started thinking about that balance between overly relying on AI and technology and the need to make sure that we treat people in a human way and that we have this human-centric leadership approach. And if we give too much of the power or the ownership to AI, then what sort of difference will that make? So I'm really interested in this topic because I'm seeing, as I say, within Zircon and BeTalent, we're really embracing AI much more than I would have ever expected. And it's making a real difference. It's accelerating our work, it's helping us be more curious, it's challenging our thinking, and it's improving our language as well. I can write a whole series of questions, put them into AI, and then it comes back and says, you know, you could improve it this way, which is just blows my mind actually how quick it is and how fantastic it is. So it's really good with creativity and efficiency, but where does that stop? You know, when should we use AI and when should we stop using it? So that whole conversation for me is really quite interesting.

Angela Malik 3:18

I'd love to hear from you, Simon, as well, about what interests you about this topic of AI and humanity.

Simon Defoe 3:24

Yeah, I think uh I build on what Amanda said really. I suppose my reflections are similar to yours, Amanda, in terms of personally, when you're using like whatever agent you're using, it's it's helping you with creativity and also doing things quickly and more efficiently. It also probably got me reflecting a little bit, you know, and this isn't necessarily a consequence of AI as such, but over the years in our space, particularly if I'm thinking of occupational psychology and assessments, we've quite often tried to turn everything into a data point to be like actually, if we can turn this into a data to make a decision about an individual, whether it's a hiring decision or promotion decision. And actually, a lot of the time we're turning what is quite a subjective judgment into a data point as objectively as we can, let's say so. So I'm not saying there's not a space for that. And I suppose it's just got me reflecting on actually, have we been trying to do that for years anyway? But AI is making us feel more uncomfortable about it. Um, and and I suppose my other reflection is just the fact that you can be using an agent or you see examples, obviously, in in the media, you know, one minute it's amazing, and then the next minute you're like, oh, this is infuriating, it's not following my instruction, it's not doing what I want. So even just understanding that dynamic of yeah, how you can use something that can feel like it's brilliant one minute but deeply flawed in another as well.

Dr Amanda Potter 4:50

A bit like humans as well, then fundamentally, we're all deeply flawed. So, I mean, it depends what our motivation is, doesn't it? So, if the motivation is around efficiency, then it does most of the time tick the box because it does make us more efficient. We're finding it's creating a whole load of efficiency for us within our business. We are definitely faster at solving problems, at writing, at doing a whole load of stuff as a result of using AI, which has been nicknamed Gerald in our company.

Simon Defoe 5:22

Yeah.

Dr Amanda Potter 5:22

So as a result of having a conversation with Gerald, we are definitely more efficient. So from an organizational perspective, it's great. But what about from an individual and human perspective? I know some of our team have really been quite reluctant to embrace AI. Now we have a closed system, we've paid so that we make sure that all of our data is safe and our learning from Gerald is safe. But some of the team have been quite reticent. What's your perception within Vodafone Simon?

Datafication And Decision Making

Dr Amanda Potter 5:50 Simon Defoe 5:50

I think there's a broader question there, which I think, particularly in HR and for organizations generally, around what is the actual aim, what is the goal, why do we want to use it, which I think some organizations will have a clearer viewpoint on that. I think sometimes we're often in conversations where it feels like adoption of AI is the goal. But I think actually what I would reflect on in terms of us as an organization, it feels very driven towards some of those things, particularly customers. So where a lot of the experiments are happening is around how can we make interactions with for the customer better to resolve queries first time. You know, quite often you don't necessarily interact with your telco provider that often. It's when you need something new or when you've got a problem. Like and actually, you know, I suppose the best service is never having to contact them at all, right? So I think there's something there definitely for us, and there's some like really great stuff actually that we've seen happening. And then I think to your point, there's a bit around definitely a drive towards efficiency. And I think there's like the learning, I suppose if you go outside of HLR and outside of organizations even start going, actually, where can just AI do it better than humans can? I don't know, maybe self-driving cars as an example, right? There might be less accidents if everything was self-driving rather than humans being involved. So that there might be tasks like that. And then the other bit which I think we're starting to lean into a little bit more, which I think maybe we didn't a year or so ago, is actually where can you create better products and services? Whereas I think 12 months ago I think the conversation was very much about efficiency and stuff like that. Whereas now I think it's actually where can it be better than humans can, or or just help us do our tasks or our jobs even better.

Dr Amanda Potter 7:40

That's an interesting move, isn't it? Because I think we've been quite short-termist in the past with AI, which is how can we solve the problem quicker? How can we get to the answer quicker? We're using it in a very short-term, immediate way. But I agree, I think that actually the application is much bigger, is that creativity and learning piece? But that's that's a really nice segue into the humanity point. Because if we're using AI to be creative, to expand, are we stifling that opportunity to encourage people to learn themselves? Because Simon, you and I have got decades of experience as psychologists. So we have spent years understanding what good and bad looks like around psychological principles. So when AI produces something, we know from our experience whether it's good or bad. But somebody who's more junior, if they were to ask AI to build something based on psychological principles, they would have less experience about what are the biases that have been included in that and how applicable is it for the client, et cetera, et cetera.

Angela Malik 8:42

Yeah, there's a risk, isn't there, that we can't assume that whatever AI model we're using knows what good is. So if you don't know what good is yourself, it's really hard to recognize where those problems might be or potential risks.

Adoption, Aims, And Customer Impact

Speaker 8:50 Simon Defoe 8:56

There's a couple of really interesting points there. And I was reflecting on this a little bit in the sense that there's been times where I might ask whatever Gen AI tool I'm using to say, oh, can you give me the latest research on X, Y, and Z topic, or I'm doing a presentation on this, or whatever it is, and asking it just to give me like references and backup support to a topic. And sometimes I'm reading the research in there going, that contradicts everything that I've learned over 10 or 15 years, and I just know that's wrong. And then I'm going back and oh, maybe I'm out of date with the research, maybe maybe it's me, right? So I'm like, okay, I need to research into this a little bit more because that totally contradicts everything that I thought was true, whilst knowing obviously things evolve.

Angela Malik 9:43

And it tends to be a hallucination.

Simon Defoe 9:46

Yeah, it's just giving you like really confident answers, and then when you tell it it's wrong, it goes, absolutely. And even when it's right and you tell it it's wrong, it goes, absolutely you're right. I am wrong because it wants to please you. So actually kind of just reinforcing whatever my views are anyway. But having that independent knowledge, I suppose, outside of the tool, I think is really important.

Dr Amanda Potter 10:08

Yeah, I think that's a really, really important thing, which is that goes back to the Jig Ramge podcast, of course, which we're going to talk about later, which I think which episode was that, Angela? Do you remember? 71, I believe. Episode 71. We were talking about big data in that podcast with Jig, weren't we? And good data in equals good data out, good interpretation, whereas bad data in, bad data out. And so we need to both be able to ask the right questions, but also understand how to interpret the answers that we're getting from AI. We need to almost to be the expert users. If we come to it without having any knowledge at all, then we will be blindsided, is kind of the view.

Simon Defoe 10:50

Yeah, I was listening to a different podcast a while back on the economist, I think it was. Yeah. How dare you talk about it? I'm so sorry. I won't name the episode just in case people gave it and not the people. Um but I suppose a the point they were making on one of these episodes around AI, and I was like, oh, that's quite interesting, was it deals with probabilities and not with facts. So when you're asking it for information, it will search based on all of the relevant information it's got and bring something back to you to say, yeah, this is the likely answer. It's not necessarily factually, this is the correct information based on what you've asked, which I thought was really interesting. Because then it obviously has wider implications when you go into sectors like healthcare where facts are much more important.

Dr Amanda Potter 11:38

Fundamental, aren't they? And it's interesting because we use Gerald for a lot for neuroscience. So we'll ask it a question around, let's make this up, but sleep and the neuroscience. What we'll then do is we'll produce it and then we'll send it to Keshire. So Keshire being our neuroscientist, she will then go through and validate everything that's come out of it. And most of the time, because it is much more fact-based with neuroscience, it it's pretty good. You know, it does get quite a lot of it right. But with psychology, because psychology is a softer science, arguably, than neuroscience, actually, there's so many disciplines and so many arguments of thought and so many papers and contradictory arguments for points of view that actually it's quite hard to use AI in some instances with psychology.

Angela Malik 12:24

It can't actually differentiate, but it can look at a healthy debate. And like Simon said, it's judging on probability and not actually sitting there and thinking about weighing up the options. Yeah, what seems right. Yeah.

Dr Amanda Potter 12:37

We always talk about psychology being the science of the obvious, because by the time you've done a study and you found the answer, you go, okay, they uh that seems quite obvious now. Now we've done it, now we've proven it, that seems right. Which of course it can't do. But this whole topic was about AI and humanity. So I suppose we should take a step back, shouldn't we, Angela?

Angela Malik 12:55

I do feel like we're way ahead of ourselves, and it might be helpful just to define what we mean by humanity. What is to be human?

Dr Amanda Potter 13:04

Humanity really is the essence of being human or the qualities that make us humane. So we're really talking about empathy, kindness, compassion, and we know those things come from a neuroscientific perspective from the vagus nerve. So the activation of the vagus nerve. And

From Efficiency To Better Products

Dr Amanda Potter 13:20

we're more likely to activate that vagus nerve when we're high on resilience, highly collaborative when we're with people that we trust and that we know when we feel psychologically safe. So if we're resilient and psychologically safe, we are more likely to express those emotions and have those feelings of empathy, kindness, and compassion because of that vagus nerve activation.

Angela Malik 13:41

I'd just add on to that the element of being authentically creative. What do you mean by that, Angela? What I mean is that the way that large language models work, they can appear to be very creative. But like Simon said, they're working on probability, they're tokenizing every request and then compiling probable answers to that. And they can produce results that look really creative, but everything that's coming out of that was put in initially. It's pulling from its data sets. So there's nothing truly creative that it's producing, it's just recombining different things that have already existed. Whereas a human could have an aha moment and come up with something truly, truly new.

Dr Amanda Potter 14:25

Simon, I think you've got a little bit of a point there around how can AI have humanity when it's got no soul, if that's all right with you.

Simon Defoe 14:32

Yeah, I was um just an additional point as well. I was on my dog walk this morning, I was thinking about this and on this question around like humanity specifically and that distinction between can AI can show those qualities or experience those qualities, even. And I was thinking about my football team over the weekend had a really good result. I was pleased with, but then I was thinking back actually to those experiences, and say, as in my childhood, I was like, my favourite footballer was Dennis Bergkamp, was like my favourite footballer. Just if you watched him play, you'd be like a lot of people would say, like, it's an artist on the football pitch, you'd watch him, it'd be really elegant. And then I was thinking about, oh, you also remember those moments of like the games you've watched with family and friends, and you recall those experiences and that connection. And then I was thinking, what would happen if I put into Chat GPT? Tell me about Dennis Bergkamp. I'd say, oh yeah, it would give me some facts, you know, he was a footballer that might be accurate as a footballer in this era, he was known for X, Y, and Z, but it can't have any of those experiences of how you kind of contextualize that. So it's so I think that was an important bit there for me on your previous point, really, about understanding the context, which AI can't do for you, whether that's in a kind of social setting like that, but also in a workplace setting as well, like contextualizing the meaning and how things might work, how things might land, etc. etc. And then I think the broader point we were talking around around AI and humanity was to the point I made earlier around actually how it can get it so right and also so wrong. I think some of the examples, which I think a lot of people would have seen in the media, you've got some great examples, we'll start there in terms of helping to detect cancer more quickly and and more effectively and more accurately. I think I listened to a a really brilliant example, which was to do with people that basically lose their speech and body functions later on in life through like MMD, and basically the the AI had captured the person's voice before they had lost their voice, and then it was capturing them blinking, and it could understand basic patterns of what they're trying to say through their blinking, and actually then being able to communicate still with family and friends via that method. And I remember listening to it, thanks. That's such a brilliant application of

Risks: Hallucinations And Expertise Gaps

Simon Defoe 16:55

AI. But then also you've got the total opposite, where I'm sure we've seen around actually people using a Gen AI tool as a therapist, and because quite often they're programmed to go, oh, I will just comply kind of with your thinking and encourage you with whatever you say. Actually, people might say, Oh, you know, I'm feeling suicidal, then it'll say, That's great, maybe we can think about a location. I think was one of the examples.

Dr Amanda Potter 17:20

It's a real case, isn't it?

Simon Defoe 17:21

Yes, so so you've got those on the opposite extreme, and obviously appreciate the people that create these tools, then try and put in guardrails afterwards. But I just suppose it it demonstrates the opposite extremes of how it can be used brilliantly versus when there's no guardrails in place, what can happen.

Dr Amanda Potter 17:36

Do you know what? That is such a brilliant and an extreme example. On the one hand, it's given an individual their voice back so that they can almost be human again. And on the other hand, it's potentially at risk of taking a life away. And actually, there was a case, of course, where someone was encouraged by AI to commit suicide and was successful. But that's quite a public example and situation, hence the guardrails being in there now, which is absolutely critical.

Angela Malik 18:04

Sort of removing and giving humanity in those two examples, right? Mm-hmm. What do you think the human agenda is right now in organizations?

Simon Defoe 18:14

Good question. Um I think ultimately for me, and I'm not sure if everyone would agree with this, but it comes down to supporting individuals being at their best and achieving what they need to, and and obviously helping the organization deliver. I also think there's an element of creating like better work and better jobs for people. So I think there's an element of where can we remove like repetitive stuff, but actually make work more meaningful and more interesting, hopefully. I'm one of the bits suspicious from thinking about this topic, because I was like conceptually, I'm like, this is really interesting, but I was also still like kind of forming lots of my opinions on this stuff and it's changing. And I think ultimately as like what we all as individuals want goes back to again, if I'm going back to the the psychology of these things, is what do we want at work? We want the sense of belonging, a sense of purpose of who we work with, we want a sense of mastery over our profession and what we're doing, so we feel competent, which kind of links back to my football example as well. And we want the ability to have a level of autonomy, right? So we're not being micromanaged or we've not just got agents around us, let's say, doing everything and and we're not really doing anything. So that for me is kind of the human agenda, really, helping people be at their best, thrive, enjoy the workplace, but also experience it in a way that motivates them and drives them forward.

Dr Amanda Potter 19:41

Well, what a great aspiration. I mean, all of those things we talk about all the time on this podcast around belonging, purpose, autonomy. And all of those things come down to psychological safety. And the foundation of psychological safety, of course, is trust and human connection. And so maybe that's the risk. Maybe the risk is that if we overly rely on agents, on Gerald, on AI, what we're going to stop doing is solving problems together. We're going to go, instead of me coming to you, Angela, and saying, look, I've got a problem, how am I going to solve this? I'll go to Gerald and just ask Gerald how are you going to solve this. And then between Gerald and I, we'll come up with the answer. And we'll lose some of that creativity that we would have had by you and I, Angela, having quite different opinions about things because we're cognitively diverse from one another. So, like we say, Gerald learns and will learn my way of thinking because of the amount of data that I'm putting in and the questions that I'm asking and the responses that I'm using because of the coding. So, therefore, to

Defining Humanity And True Creativity

Dr Amanda Potter 20:45

what extent are my biases being imported into Gerald? And is Gerald therefore feeding back my biases back to me? And if we think about decision making at a senior level and our decision-making questionnaire, one of the things that we try to really do is look at decision dynamics and encourage teams to think about in what ways do the team come together collaboratively and what ways do they disagree and challenge one another. And we want the challenge more and more in order to drive increased problem solving, increased innovation and creativity. And I'm slightly worried that if we go to AI, then we're all we're doing is going to be confirming what we thought already because it's going to be learning what we think and we'll be sending back the messages that we want to hear.

Angela Malik 21:30

That's why I often ask it to critique my thinking. I'll give it my original thinking without just asking it straight out for its first thoughts, right? And then say, critique this and tell me what do you reckon? Very good. January can feel like a fresh start, but it can also be a low point for resilience. Blue Monday may be a fabrication, yet it reflects a pattern we see every year. People return to work with high expectations and limited energy, and many. Leaders feel the strain. That's why BeTalent and Kenneco are hosting a live Ask Me Anything on Resilience on January 19th. We'll explore what current research says about resilience, why it matters for performance, and how leaders can create the psychological and behavioral conditions that help teams navigate uncertainty with clarity and confidence. If you would like to join the conversation or ask a question, visit the CPO.co.uk to register.

Dr Amanda Potter 22:30

So, Simon, what do you think the biggest risk is about the use of agents or AI within organizations? What do you think it is from the Vodafone perspective?

Simon Defoe 22:39

I think it's broadly in line with what you're saying, but whether it's going to exacerbate that kind of bias that we've got as humans, bias in terms of the build of those tools themselves. So actually the fact that you can go into without me being an expert in this topic, but you can go into all of the different AI tools, you could ask exactly the same question, and they can give you quite different answers and output. They do also shows you in terms of the philosophy and the approach of the people that have built them, have all got very different stances. So I actually think you're almost shaping the direction of your organization potentially by your tool of choice. So I think there's an interesting bit there for me around the guardrails and how organizations approach these things. I think that's where I'm at with my opinion on it.

Dr Amanda Potter 23:29

To what extent are you training people to use AI in your organization? Because if you think about emotional biases shaping how we ask questions and how we interact with each other as humans, and our resilience has an impact on our propensity to be in control of our personal biases? So if we're low on resilience and low on psychological safety and we're feeling frustrated, feeling annoyed, our language is going to change and the questions that we ask are going to change. So, to what extent are you training people how to interact with AI, how to ask questions, and how to make sure biases don't shape their emotions or their behaviour or choices and the way in which they use it?

Simon Defoe 24:12

I think what's quite interesting, I suppose this is a broad comment rather than for us specifically. I think a lot of our training and a lot

Context, Meaning, And Lived Experience

Simon Defoe 24:20

of organizational training is around the technical aspects, quite often, like how to actually build your own agent, etc. etc. There's an element that we're pretty good at as well in terms of how to write a good prompt. So making sure like you've got the context, what sources you want it to search, what the expectation is in in terms of your asks. We also encourage quite often what you were saying, Angela, and I use it in the same way as you actually to go give my thoughts first and then say, ask me some more questions to get a bit deeper into my thoughts, and then I want your critique afterwards. So we very much do stuff around that. But I suppose just the nature of this conversation, and I've not really seen this externally either, is how do you really get people to supplement their judgment with AI? Like so you're using those outputs, and how do you really keep that human judgment, that creative challenge, let's say, in kind of essence of what you were saying, Amanda, that actually it's sometimes it's good to have meetings with groups where you all come at it with from different angles, you put your positions out on the table, you have that healthy debate, and it's okay to disagree on things. Um, so that maybe feels like to me there's there's maybe a little bit missing there in terms of that real, let's say, like AI human interaction and collaboration, let's say.

Dr Amanda Potter 25:49

I think there's something to do with accountability and ownership as well for me. Because if there's a big decision that needs to be made in a boardroom or in a leadership team, the debate and the decision making means that people start to take accountability of the decision and the outcomes. I think the issue with overly using an agent, Gerald AI, whatever we're gonna call it, is that there's that risk of external locus of control. Well, it wasn't me, it actually was Gerald who came up with that answer. So if it doesn't work, don't blame me. You know, we can try it, I'm happy to try it, and that's what AI says is the right answer, but you know, I'm not gonna completely commit to it. And that's potentially quite a big risk, actually, because you could find that there's a lot of flipping and flapping around, depending on the problem, rather than actively agreeing the way forward. So that's something for me that I think is potentially a risk, which is around that ownership piece.

Simon Defoe 26:46

And it's almost where organizations thrive and people thrive, is making those bold decisions that AI won't be able to help you with. It might be able to give you some ideas, just like big data can give you, or data itself can give you some guidance and some insight. But ultimately, it's still an individual that has to take the risk with that decision. So you you could use AI as an example, right? Actually, the people that 30- 40 years ago started exploring actually, oh, I think machines can maybe think in the same way as humans, their idea of their research topic. At the time, a lot of people were like, that's not a good research topic, that's not a thing, that's a but actually, you know, they had this idea, they stuck for their guns, obviously, and you know, they've founded this same way. Obviously, you can think of some of the AI companies that have been founded. And also it made me think of the episode you guys were referencing on the data

Life-Giving Uses And Harmful Failures

Simon Defoe 27:45

part. There was a part in there around the stock market, and it was saying that actually if you went back however many years and said on the traditional measures of how you would decide to buy a stock in terms of their earnings, their debt, their growth revenues, with a lot of tech stocks now, all of those fundamentals don't exist. It's totally overweighted, almost you could argue, a bit of a bet that it is going to work out and they are going to make the revenues expected. But people are having to take those bets to say, actually, this is worth the investment in time and money because we think we'll we'll get there in the future. Whereas I'm sure if I ask ChatGPT what tech stock should I invest in or whatever, it would say, Oh, these are the ones. But actually, ultimately, it there's a human judgment in there on whether that is the right risk to take or not.

Dr Amanda Potter 28:33

And it can't predict the future. So with something like that, it can with a degree of probability again talk about the past, but it can't predict the future, nor can we. But we have a good idea of what's going on in society and in the environment and in organizations so that we can start to have an idea of what might be coming ahead, which we start using all that information. So I suppose that's a really good point, which is AI can only ever interpret what it's been given and what's happened so far. It can't think forward. And I think that's one of the things that as humans we're very good at and thinking about the potential, the so what what might happen.

Angela Malik 29:15

I think I would challenge that a little bit, Amanda, because I think it can project, again, based on probabilities. But if that projection is based on any wrong premise, and there aren't those human nuanced aspects of consideration that get thrown in there. Unless the prompt is super specific about that or something, then yeah, it can't predict the future. But in certain areas, I think it could project quite accurately if it had the right data set and the right prompt and the right and and and.

Dr Amanda Potter 29:45

So I think this is be a nice time just to summarise a lovely piece of work that was done by Jess Ross in our team. So she did a whole article all about psychological safety and AI because I think it's really relevant with humanity. And what she did is she did a whole analysis of the connection between uh psych safety and AI. And she came up with a few points that I thought were really useful. So if it's okay, I'll just quickly summarise them, which is uh when we're thinking about the two aspects, we have to remember that humans we seek psychological safety meaning connection. AI does not. Number two, as technology changes, uh this can trigger in us humans a fear and uncertainty which can undermine psychological safety. But AI is not worried about that. And number three, in order for us to thrive and strive and grow and learn, we need to be creative, we need to be emotionally intelligent, we need to be resilient. AI does not. And all of the things that are important for us to succeed as humans are all wrapped up in humanity. But AI does not need that in order to survive and thrive. So it's just really interesting to me that the thing that will fundamentally help us as humans, as a human race, uh succeed is humanity to be kind, to be compassionate, to be empathic, to be good. Yet the one thing that AI needs is not that, actually, it just needs data. So it's slightly worrying and slightly scary for me. I have a vision of us all walking around with tiny little pinheads in about three decades because our brains have shrunk so much that our heads have shrunk. So we walk around, we've got heads like what movie was

The Human Agenda At Work

Dr Amanda Potter 31:30

it? Beetlejuice. Doing Beetlejuice and the head shrunk, the man is head shrunk and he's in the waiting room. I have a vision of all of us having teeny tiny heads in the future because we've lost our brains. We no longer know how to use our brains.

Angela Malik 31:43

But Amanda, no, because we know from the neuroscience that when our threat perception, when our cortisol is high and our threat perception is high and sustained, that our amygdala actually grows. So our brains won't get smaller. We'll just have a giant amygdala. But we just have a giant amygdala and our prefrontal cortex is tiny.

Simon Defoe 32:02

I thought you were going to go down a totally different route there. That's what Amanda was gonna say. Like you're walking around with our own like humanoids, and we're walking around, so you've got Gerald the robot with you coming out or something.

Dr Amanda Potter 32:14

With my tiny head on my shoulders, a little pinhead. So yeah, I kind of worried. The reality is it does quite a lot of the thinking for us, and we need to still remember how to think. We still need to apply ourselves and not just take at face value what we get out of AI.

Simon Defoe 32:32

I do think there's a risk there, because I find myself doing this at work sometimes, because it helps you do things quicker. You're sometimes then not putting the real thinking time into creating a proposal or the output. And sometimes that's fine in the context to your point earlier, Amanda, when you know you know you've done some of this work for decades, and actually you can quite quickly see some outputs and be like, yep, that's really high quality, or that's not. But sometimes when you're doing something new and you need a bit of the thinking time and and you need to know, like shape your own thoughts and opinions on a topic before being able to go with a proposal or an idea. Actually, sometimes it gets you there so quickly you haven't gone through that process, which can be super, super important on I suppose any topic that you're working on.

Dr Amanda Potter 33:22

I completely agree. I'm working on a project with Caitlin at the moment, which is a brand new piece of work, a brand new area, one that we have published a podcast on, but it's an area of expertise slightly outside of what we do on a day-to-day basis. And so, therefore, the ideas and the solutions we're trying to generate from the data are completely new. And I did put into Gerald, you know, what would the typical response be to this problem? And it came back with a really good answer. And then I kind of read it and thought, oh, I'm not sure that's what I would do in this scenario. So I am applying the thinking in that situation because it is brand new. I'm using it to challenge me rather than to guide me. So I think you're right. I think when it's a brand new topic or a new area that we need to really explore and understand, it can be helpful, but we shouldn't overly rely.

Simon Defoe 34:10

The more I'm like hearing us, the more I'm thinking it we're emphasising, aren't we, that like critical thinking, that judgment, that ability to be creative and critique things, as well as all those other things that feel very, very uniquely human.

Dr Amanda Potter 34:24

Yeah. And the risk I'm really worried about, which is what Jess Ross was worried about, is that if we overly rely on AI, then are we going to diminish that capacity and are we all going to become pinheads?

Simon Defoe 34:37

Now I'm picturing both me and you of our pinheads talking and you this time.

Dr Amanda Potter 34:46

I've just had those bubble head things in my inside mind. No, mine's definitely just the pinhead from Beetlejuice. Actually, I mean, this is a great topic. We should carry on talking about this stuff in all of our pods. You know, the reality is it's changing the way we work. It's the one thing, other than the internet 30 years ago, it's the most fundamental in changing the way we work, I think, for me.

Overreliance, Bias Loops, And Challenge

Dr Amanda Potter 35:10

It's pretty revolutionary. It's speeding me up. But what's the downside?

Angela Malik 35:15

So I guess final question from me how do we make sure that the future of work is not just technologically advanced, but also deeply humane?

Simon Defoe 35:24

I feel like I could talk about that for a long time. I'm actually, although the conversation I think so far has been about some of the challenges and the uniquely human things, I'm also very optimistic we can improve the quality of work for people, for individuals, and for organizations as well. But I think that also needs a clear steer from organizations and probably from governments actually, in terms of how should AI be implemented, what's our purpose, what's our goal with it, and how do we expect individuals to work with all of this tech? So I quite like if you think of some of the things you're seeing already. So I think some of the pharma companies is at Moderna that have merged their HR and IT function as an example, because they really understand that actually the future of work is going to be even more HR and tech integrated. I was also thinking, good luck to the search firms trying to find like senior leaders that have both the tech and the HR skill set, but maybe that's where even more looking to.

Dr Amanda Potter 36:31

Because it's the future.

Simon Defoe 36:32

Yeah. So I think I'm quite optimistic in that sense, but I do think there probably needs to be a bit more from organisations and from us as individuals about what is it uniquely that we bring. And I think we've discussed all of that. And I think from there you kind of understand where the the tech sits. For me, the tech is around like reducing the boring, mundane task, let's say, that that can be done quite quickly, but also then elevating that kind of creativity and yeah, critiquing kind of the human side of work as well.

Angela Malik 37:05

So I think that brings us to the end of the episode. It's been a really fascinating discussion, and I think there's going to be more and more conversations that follow after this because it's such a broad and comprehensive topic, isn't it? AI. But in terms of how humanity and AI come together, we've seen examples where AI can take our humanity or even give us humanity. But we can't forget to be conscious of those psychological safety risks, as we've talked about on the pod here. And we have to keep in mind how AI actually operates and what inherent biases it might have. Remember to still have that challenge in our mind and think clearly about what the output is that's coming to us. But we can still be optimistic about the future. We just need to, as leaders, as organizations, address those risks and adopt that technology with human-centered leadership and approaches in mind; would you agree, Amanda?

Dr Amanda Potter 38:03

Yeah, I completely agree. I think that we just need to keep an eye on it, not overly rely, make sure that we use it ethically and continue to have conversations with our colleagues, not just only have conversations with Gerald.

Angela Malik 38:17

I think that brings our episode to and end. Thank you so much, Amanda and Simon, for your time and thank you to our listeners. If you like what you heard, feel free to give us a rating so more people can tune in to discover all of the fabulous insight that our lovely Dr. Amanda Potter brings every episode. Thank you so much.

Dr Amanda Potter 38:33

Thank you, Angela. Thank you, Simon, so much for being our guest. And thank you, everyone, for listening. You have a wonderful and successful day.