#191 How Generative AI is Changing Business and Society with Bernard Marr, AI Advisor, Best-Selling Author, and Futurist

Hi, I'm Maggie, VP of Curriculum at datacamp and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

There will be some job displacement for sure, but the vast majority of jobs will be augmented and where AI will almost give us a superpower that works alongside us and will make us better and more efficient and hopefully more human in the long run. And what people should be worried about is not AI, but people using AI better than they are in order to become better at their job.

Welcome to Data Framed. This is Ritchie. Pretty much everyone agrees that generative AI is having a profound impact almost everywhere. The tricky part is that thinking you can do anything with generative AI isn't that helpful to deciding how you should make use of it. So today we're going to focus on specific impacts of generative AI in some key areas. We'll discuss how generative AI is affecting employment, learning and upskilling, writing code and engaging your customers. Joining me today is a true data and AI celebrity. Bernard Maher is the best selling author of 19 books including Generative AI in Practice, which forms the basis of our chat today. LinkedIn recently ranked Bernard as one of the top five business influencers in the world and he advises many of the world's best known organizations such as Amazon, Google, Microsoft, IBM and Toyota. I have a small confession to make that back when I was doing curriculum design for datacamp, rather a lot of my course ideas were inspired by things I found in Bernard's books. That means I'm pretty thrilled to finally meet the author and to pick his brains. Hi Bernard, welcome to the show.

Hi Richie. So lovely to be with you.

Excellent. So I'd like to start talking about impact just at a high level. Where do you think AI is having the biggest impact on businesses?

Yeah, this is a question I get asked quite a lot and my answer said it will actually impact every single business in every single industry. There are businesses that will be affected more in the short term. Typical examples are customer service industries where we have chatbots replacing what they do. But I feel that pretty much every organization needs to really think hard about how AI will transform their business. And it's going to. The reason I wrote this book and I'm so excited about this book, is that it is really the most transformative technology that I have seen in our lifetime that will impact everything. And from a business perspective, what they need to do is they need to look at their own products and their own services and think, hey, how will generative AI and AI in general transform this? How will make it, how will it make our services more intelligent, our products more responsive, more personalized, and how, of course, it will transform their own operations. So there are huge opportunities around automation and augmentation and driving efficiencies in businesses. And I truly believe that there's not one company and not one job on the planet that will have, no, will not feel any implications of generative AI.

Alright, so affecting everyone and lots of different use cases there. So we've got our work cut out for us in the next 45 minutes.

Or so we could. Top of 45 days.

Absolutely. So just beyond business, what do you think the impact of generative AI are for wider society?

Yeah, because it will affect every industry and every business. It also will affect every single job. So this is an impact that we will all feel. And sometimes there's a lot of scaremongering out there. The magazine and newspaper headlines that drive traffic is okay, we kill all our jobs and make you jobless. There will be some job displacement for sure, but the vast majority of jobs will be augmented and where AI will almost give us a superpower that works alongside us and will make us better and more efficient and hopefully more human in the long run. What people should be worried about is not AI, but people using AI better than they are in order to become better at their job. And in terms of the impact on society, I have concerns about are we ready for this? Is the world ready for this? Are people ready for this? Are we prepared for this? Can we actually transition into this new world? And have people got the skills that they need to really leverage AI in their jobs? And that is a concern I have. Do we really have people that are willing to retrain to rethink their job and their career? And especially when I look at some, some lower level jobs might not require a huge amount of, of cognitive skills for some of those people, I have serious concerns about how they would be ready and thriving in a world where AI will become so dominant.

So it seems like there are two stories here. There's the happy story where people have their jobs augmented by AI and become more productive. And it's brilliant. And then the scarier story about where AI is displacing jobs and people then do the jobs, it's like, well, you gotta do something completely different. So do you have more of a sense of who needs to be worried? Like, what are the jobs that are most gonna be displaced? And are there any skills that are more resistant to being taken over by AI?

Yeah. So there are lots of different studies at the moment flying around. I was at the World Economic Forum in Davos, and the IMF presented a big study they had done. And what they found is that 80% of jobs in the developed world will be affected by AI over the next five years. And. And this goes down to 60% for the rest of the world. So jobs that are actually quite safe in terms of augmentation are the jobs where we need some dexterity. And so, for example, if you're a plumber, that is pretty safe. If you're a hairdresser, that's pretty safe. If you are working in nursing and healthcare, that's pretty safe in the short run. But all jobs that have a component that is repetitive, that is easily automatable with current AIs, and also the creative jobs are changing now. And I think this is where I see a huge transformation that in the past, if you were, let's say, a songwriter or a writer or someone that composes music, you will be you with your instrument, you with your writing pad. Now you have almost a copilot with you. That is AI. We can bounce off ideas. So if you are a product designer, for example, we now have generative AI software. So if I wanted to design a new car or a new anything, you can bounce off your ideas and you can say, this is what I want. I want this weight, I want these characteristics. And the AI can then come up with new ideas. And this is, for me, really fascinating. I've worked in pretty much every industry, and I recently worked with the whiskey industry, where they're now using generative AI to come up with new blends of whiskeys. Because this is a really complicated task where you have lots of existing barrels of whiskies that have all been aged in different barrels for different length of time. And they to come up with a really meaningful commercial offering that's difficult, that also tastes good. And again, they now work alongside AI, where they look at the stock levels, they look at the commercial realities, they look at what consumers actually like, and then come up with new recipes. And then the blending master sits there and tries them out and fine tunes them together with the AI. And I think this is where the sweet spot is. This is where I'm particularly excited about. And in terms of jobs, it means any. I think it's hard to say job specifically, but more if you almost break down your own job and say, okay, these are the tasks I perform every day. As long as you understand the capabilities of generative AI today and in the near future, you can then make an assessment, how much time do I spend doing these different tasks and how likely is it that AI will be able to do them in the future? And then you try to focus on those that AI can't do as easily. So for me, in terms of skills that we need to concentrate on as humans are the ones where we don't compete with machines that are very good at certain things like analyzing data. So if you're a data analyst and your job is to extract low level insights from data and turning this into reports, this is something AIs can do much better than humans. They can analyze unstructured data, huge, vast amounts of data, analyze the sentiment of both structured and unstructured data, analyze it and really change the nature of analytics from one when analysts analyzed information, packaged it for the people and send it to them, to a push concept where the AI will analyze data and say, hey, looks like there's an interesting insight here for you to look at, where you then can have a conversation with your data and where the reporting is completely customized and automated to your own preferences. So some people might like a text report, some might like a graph or a table, and the AI can give give you all of this. So if this is your job, I will be worried. But if your job involves critical thinking, which higher level analytics involves, emotional intelligence like in nursing, like in teaching interpersonal skills, or the physical dexterity that I talked about, then your parts of the jobs are much safer for the future.

Wow, there's a lot to unpack there. And I found it very interesting that you're saying that things involving dexterity, they're going to be very safe from AI for a long time. So maybe initiate that backup plan to become a hairdresser.

I would probably caveat this slightly in the long run, in the medium term, definitely at some point I have concerns that if we are merging the generative AI capabilities with robotics, that we might enter a new stage where robots will be hugely capable and will be able to do things like hairdressing and bricklaying and plumbing, but we are still a bit off that, absolutely.

So the robot hairdresser is kind of interesting, but yeah, not coming anytime soon. So you also mentioned critical thinking, and I'm wondering, does the existence of AI then mean you should change your approach to your career? Like, what should you be doing differently? And I guess, does it matter if you're like just starting out your career, like you're 20 and you're trying to figure out what do I do with my life? Or if you're 50 and maybe you sort of late career stage?

Yeah. And I mean, I have three children between the ages of 12 and 18, and this is a question I'm thinking about a lot because I want to give them advice. And I think what it will mean for our careers is that careers will be much less predictable because the pace of change is going to accelerate. If we think we've seen lots of change over the last 10 years, that would be nothing compared to what we'll see over the next 10 years. So traditional careers will change, they will be much shorter. People will change jobs more frequently and really rely on the truly human skills that they have. So if you have your critical thinking skills and emotional intelligence, I actually wrote an entire book called Future Skills that looks at the 20 skills I believe people need to be ready for the future. And there are some technical skills. They're skills like, okay, I need to understand the AI capabilities and I need to understand data to some extent and have some data literacy, and I need to understand some of the cyber threats that we are now facing. But beyond this, they are all truly human soft skills. They are about our ability to make complex decisions, our ability to have creative problem solving. All of these things will be skills that will enable us, as long as we leverage AI alongside those skills that will allow us to thrive in the future.

That's quite nice that a lot of it really is. It's human skills. Place your human strengths rather than trying to out compete the machines.

Absolutely. And this is probably my biggest recommendation. Wonderful.

And so learning new skills seems to be a very important thing, and this is something we worry about a lot at data camp is how do you get better at stuff? So how can AI be helpful for people trying to learn new skills?

Yeah. So one area that AI is completely transforming is education. So AI can help us learn because it can understand what we know and it can then customize and personalize the teaching to us. And just imagine you like, I don't know, you want to learn about data science or you want to learn about critical thinking, and you have your particular learning Style. So you might be a younger person that loves a certain celebrity and they love rapping. If you then wanted to learn this, we can now use generative AI to personalize the content to you and then personalize the delivery. So if you wanted this celebrity to be your teacher and half of the curriculum is delivered in a rap, you can do that. And this for me is super exciting that it will help us. And for me, in terms of what we should be fostering in US is there are two skills or two mindsets I think are particularly important. One is curiosity. So this ability to be excited about the future and wanting to learn and explore and understand the risks and so on, and humility, where we actually step back and say, actually this is completely changing our world. I don't quite understand all of this yet, but I need to understand that it is going to transform the world. And so if you are curious, if you have humility, then this allows you to have this flexible approach to life and this thirst for lifelong learning.

Curiosity and humility, just two very important traits. On the flip side to this, how do you think AI can be used to help teachers?

Yeah, so I, I've just done a, a few videos on my YouTube channel on the impact on of AI on education. And on the flip side, it would be hugely helpful for teachers. They're already using it. There was a, a recent YouGov poll I think in the UK that found that 60% of teachers are already using it today. They're using tools like ChatGPT and to help them with creating lesson plans. It helps them to customize content. If you wanted to create a quiz for your class, you can ask GPT or Google Gemini to do this for you. And there's a lot of work going into helping teachers with their administrative burden as well. My wife is a teacher, so I know that a lot of evenings and holidays are spent marking things, preparing for things. And again, AI and especially generative AI can help us with this. It can help you mark papers, it can help you with correspondence with parents and students. It can help you analyze data. All of this will make the life of a teacher, hopefully better.

That's amazing. I'm sort of nodding vigorously here. I used to be secondary school math teacher and I have to say marking is just one of the bane of teachers existences. So yeah, having AI help out with that, automating all that way, that would be just a dream for many teachers.

Yeah, and it's interesting and I think teachers really need to figure out how important AI will be in Education and the impact it will have because one of the things that I, I often talk about is that I see this in my own kids. They use Snapchat as, as one of their favorite tools to communicate with their friends. And a few months ago Snapchat added an AI friend to everyone's Snapchat list. So suddenly a new friend appeared. And this friend is this all knowing AI that is powered by GPT4 that can answer any questions. So quite often it's like I was sitting down with my 12 year old, he had some math homework and there was a complicated formula and even though I studied some of this at uni, I couldn't quite remember how to solve it for what they were looking. So I again out came the app. We took a picture of the question. So can you please explain how we do this? And the AI simply solves the equation and explains in really meaningful detail and then you can have a conversation. I don't understand this. Can you make it simpler? Can you explain, can you give me different examples? How does it apply to the real world, blah blah blah. And this is fascinating that it enables kids to learn and feed their curiosity. On the other hand, I think in education we have to be careful that they can also say hey chatgpt, write my homework essay for me. And we need to think about what that means for education in the future.

It's a double edged sword really. You can get that interactive fun experience to help keep people engaged. It makes it much more difficult to accurately test people because it's much easier to cheat.

Absolutely.

I'm wondering do we need to change what gets taught? Like should the, should school or even adult training curriculums change because of the existence of generative AI?

Yeah, I think this is a really good question and the answer is absolutely. So if we are seeing, I'm also a governor at a big secondary school and what I see is that the curriculum is no longer fit for purpose and I don't think it has been for a long time. So if we train our children to retain facts that many of them are completely irrelevant, that is not the best use of their amazing potential if we think what humans are capable of. So for me, what we need to do is we need to make schools into these maker places where we really focus on things like creative problem solving, where we work on teamwork, working in diverse teams, critical thinking, all of these things are super important. And we need to understand the AI is so important that actually teachers have and schools and the curriculum has a really important role to make sure that our kids are learning to use these tools well and understand some of the shortcomings of these tools. So they are becoming more technology literate, more data literate, but have the human skills that will allow them to compete in the future.

That seems a really interesting idea. The idea that there should be more teamwork in schools and more focus on critical thinking rather than just fact memorization.

And if you think about it, it's like the. I don't know that some teachers might really enjoy the repetitive teaching or certain curriculum elements, but I know that I have, if I think about my own teaching, been teaching at Cambridge and Oxford and do lots of seminars and keynotes. Some days I feel I'm really good and I can bring things across really well. And there are other days where I think, okay, that was not my best of teaching. And what generative AI will allow us to do is to give children access to the best teachers at their peak, delivering any content personalized to them. So this teacher at the front of the class delivering content for kids to remember I think is out of date. They can get the input from the world's expert if they want to learn about AI. But the most amazing example personalized to if they have any educational needs, all of this can be taken into account. And a class teacher can't really do this. You try to differentiate your teaching to the different levels in the class, but you can't do this for 30 children in your class. AI can do that, and then the teacher can then play this really important role of facilitating the learning, going into more depth, letting kids explore this work and team, make things, solve problems creatively, and fulfill another important role, which is the safeguarding role. I think this is sometimes overlooked. We think we can just pass all our education over to robots. So I think that is super exciting and will make teaching better and more exciting, hopefully for everyone.

Absolutely. So just on your point about sometimes you have good days and sometimes you have bad days, another guest on the show talking about doctors and how AI can now get sort of better patient interaction scores than real doctors because it's trained to be always nice to the patients and doctors sometimes have bad days. And it seems like there's a similar situation with teachers where, you know, teacher's having a bad day, maybe they're just like, just read the textbook for half an hour.

But yeah, and for me, this is a, it's a, it's a super important example that I have a whole chapter on, on healthcare in my book. And for me there are so many examples. There's quite a bit of research that shows that mental health support, for example, done by AI is more effective because the patient feels less judged, because they know they don't speak to another human being where they feel, okay, they might feel, I've done something wrong, I'm stupid, I'm this, I'm that. So they tend to be more open. And for me, the really interesting idea is to have almost an AI that is facilitating a consultation between a patient and a doctor. Because if you think there are so many things wrong with our current healthcare system too, the heart to get a doctor's appointment. When you have an appointment, the doctor has 10 minutes to see you. It's very unrealistic to think that before the doctor sees you that he or she has had a look at the entire your medical history. They know exactly your. Your family, circumstances and whatever. So what usually happens is you describe some symptoms, they try to figure out what the problem is, and then they try to enter this trial and error period with some sort of treatment, medication, whatever. For me, there's a role that AI can play because the AI can understand you. You can talk to the AI about your symptoms even before you see a doctor. The AI can understand your entire patient history. And then when you go in there and the doctor listens to you, the AI can then help the doctor diagnose what might be wrong with you. So the doctor might say, I think is depression. And then the AI might say, no, maybe it's not. Maybe it's stress because I've recently had a change in job and this change and this and this and this. And then again, once the doctor then has made a decision or a diagnosis and the patient believes that this is the right one, then the AI can help the patient to understand what this now means. So what does it mean? I now have to take these tablets and I need to change my lifestyle and I need to do that. And that for me is super interesting that you almost have this AI that sits between a doctor and a patient, almost as a. As an observer, as a co pilot. Again, I like this word of co pilot. Or actually when you look at AI and in my book, I abbreviate generative AI to gen AI. And for me, that almost reads like a genie. So almost you have your a genie on your shoulder or you have your little intelligence, your magic lamp that you can rub genie comes out. You then work with them. And I think this is what the biggest offering that generative AI can have.

I like the idea of a genie granting you some wishes. There. And it just seemed. I really like the point you made, that because AI doesn't judge you, you can be more open with it. I think that's very important in a lot of particularly mental health situations, but in life in general, maybe career situations, you want to ask it about something or relationship situations. Lots of use cases for this. Now you mentioned the idea of a copilot and this is something else that is very important for people writing code. So can you talk about how generative AI can help people write code?

Yeah. And so for me, I've just done a video on this, on how generative AI will change the role of programmers and people that write computer code. And it's interesting that even five years ago I would have said, okay, coding is a really safe job to go into five years forward to where we are today, where Even tools like ChatGPT and Google Gemini can generate code in pretty much any computing language to a really professional level. And we now have the AI integration into GitHub. They use Chat GPT or GPT4 as their coding tool. And Microsoft, who owns GitHub, basically. And GitHub is obviously the largest open source platform for computer code. So anyone writing code quite often goes to GitHub to then look, find bits of code that they can use and reuse. And we now have the generative AI component as part of this. So for me, this copilot word or this genie that works with you is a great example in the coding world, where the job of the coder is much less about writing code and much more about actually really figuring out what do I want to do with this code, what are the requirements? And then the AI can help you write the code and it can help you test the code and run simulation. So it changes the job of a software developer in terms of instead of spending more time coding, you can spend more time gathering requirements, testing it, using AI, collaborating with your stakeholders to make sure you understand what they want. And for me, to some extent, this democratizes coding because not only will it help programmers, but I can go to tools like ChatGPT, say, please write me an app and it will do this for me, or create me a website and it will do this for me.

It seems like it's changing a lot of different areas of software development or writing code for data science, for example, I'm wondering, do you have a sense of what a new sensible workflow is if you've got all these AI tools? Like, how is it going to be different to what it is now?

Yeah, for me the key is that AI is almost part of every bit of that workflow. So if you think about this, the first step is to figure out, what do I want the code to do, what are the stakeholder requirements? And again, you can then create something, but the AI can help. You can then say, okay, am I missing anything here? Does this all make sense? Then the AI can draft the first bits of code for you, and then you can again fine tune it. You can then use the AI to test the code and say, is it working? Are there any mistakes in there? So for me, it's really important that you can automate all these repetitive tasks. It can make you much more efficient, but it won't replace some of the nuanced decision making that human. Human developers bring to the table. So very similar to teachers and doctors and coders won't be replaced, but their role will change and we will move away from the, okay, hacking code into a machine to the much more strategic, the more creative. And that for me is again, exciting.

Absolutely. And I like the idea that it's really about making critical decisions again and like deciding what you actually want to create.

Yeah, all the things that very often fall to the wayside because people are too busy coding.

You also mentioned that one use case is if I want to create a website, then I can just say, okay, go and create this. And it seems like there's a lot of potential here for people who aren't programmers who don't have that background to be able to do things involving coding. Can you just elaborate a bit more on how you think people should get started with this? Like, okay, I want to try coding task. What do I do?

I think the best thing is to try it out. And for me, ChatGPT is still the go to Google Gemini. It's another tool we can now use. I remember recently I was, I can't exactly remember what I needed to do, but I needed to do something complicated and I thought I can create a really complicated spreadsheet for this. I had some data set that I need to transfer laid into something else and then add something and subtract. So I thought I can now spend a bit of time creating a macro in Excel to help me do this, or I could just write it all down and spend 20 minutes on a calculator figuring it out. And actually I thought, let's give it to ChatGPT and see if it can create me a little program. And it did program me little HTML template that I could then use and it did the job perfect. So for me, just try it out. And for me the really important point is that this is a, again, you're working alongside the AI, so you need to learn to give the AI the right prompts and you need to fine tune it. So if you say design me dating app, it might come up with something, but you need to give it a bit more information. Or if it comes up with something, say, I don't like this, can you make a bit more of this and less of this? But it can do the job and increasingly.

So I love that approach of just try something and see what happens. I mean, it doesn't work. It doesn't work. That's fine. You're going to learn something anyway. Exactly. Excellent. So I want to go back to what we were talking about right at the start there. So we were talking about the impacts on business and it seems like one of the big developments with generative AI is there are chances to improve the customer experience. So can you just talk about what are the most common business use cases in this area?

Lots of different use cases and one of the areas where lots of organizations are using it at the moment. So typical example are chatbots. So you have a customer service team that sits and answers calls or answers web queries for you, or text queries. AI can now do this really effectively. And companies like Octopus Energy, for example, did this. They now have an AI that handles the work of 250 people and it handles almost 50% of all their customer queries. And what is interesting is that the rating, the satisfaction rating at the end of these customer service engagements are higher when the AI deals with it than when a human deals with it. So there are so many different examples that I cannot talk about how different organizations, or JP Morgan Chase, for example, created a personalized financial app that basically gives you financial advice. So it helps you create new products, different services, and the customer service that goes with it and this ability to make it completely personal. I think this is super exciting bringing all of this together. I was recently talking to a customer that works in the maintenance of washing machines, for example. And what they basically need to do is that they send out an engineer to fix your washing machine. They need to make sure that the right parts are on the van when they turn up to then solve your problem at the first visit. And if they have to have multiple visits, that reduces their profit margins or even makes it unprofitable. So it's really important. In the past, their call centers, they had some understandings of, you say my washing machine's making this noise and this has happened. And then experienced people think, hey, I think it's this part that needs changing. And they can. They run AI on this. And the AI is now a lot better at predicting which parts are required, which makes the whole organization more, more profitable to deliver better service. So for me, any aspect in, in terms of customer service will be improved and can be improved using AI.

That does seem really interesting. The idea that actually one of the biggest bonuses you can get both of the customer and the business is just making people have less interactions with the support people. So rather than taking four different visits to fix your washing machine, you get it done in one and then you got that faster turnaround time and cost you less.

Yeah, exactly. And I was recently another client of mine works in the motor insurance industry. And so if you have an accident, the whole complexity of the supply chain where you have insurers and repair shops and car companies and everyone involved, and the AI can actually streamline a lot of this. They can do the negotiations in the background and inform the person that has the insurance with this organization what is happening and make it a lot more predictive. And I think this is, that is always a good thing. If you say, okay, it looks like it's going to take another three weeks, but we're doing this and this and this and this is what we've been doing. As long as they keep you informed and you feel that this is all happening, you have a better customer service experience.

Oh man. It's back to like, yeah, keep people informed. This is communication skills again. This is soft skills are winning. All right, so you also mentioned personalization and how well Gem Jev, AI has just made personalization cheaper so we can expect more of it. I'm never quite sure what level of personalization is a good thing or what are the big use cases here. Can you just talk me through some examples again?

Everything. So we talked about education, for example. So if you use a language learning platform like Duolingo, for example, and we all have different reasons why we might want to learn a different language. So one person says, I've fallen in love with someone in this country and I want to have a relationship. The other says, I'm a chemicals engineer and I want to know work in this company that is mainly French speaking. So I need to learn a particular curriculum. In the past, language learning was pretty standardized and if you needed, the chemical engineer would have real difficulties picking up the right vocabulary. Now, Duolingo has integrated generative AI capabilities into their tool. So even Though they might have had some of the vocabulary in there, the dialogues were not as relevant. And now it can create those dialogues on the fly to me. So if I want to learn about what's the right vocabulary, what the right language to use in French when I have a business meeting about chemical engineering, and that for me is super fascinating. That is simply the tool had lots of those capabilities already. But with generative AI, it can completely personalize your learning experience if you think about just basic things, how you consume information, how you read the news at the moment, your app that you use, or you. Most people read their news online now, but in the past, a newspaper arrived, you had absolutely zero customization. You might pick a newspaper that is more in your political view or not, but that was it. And then you had the content delivered to you in a standard format. Nowadays, the app knows a little bit of what you're interested in and personalize it. In the future, it can not only personalize the content, but also how the content is delivered, the content of the content, actually. So not just this is a news article that someone has written that's delivered to you, the news article will be customized to you. So for example, when I read an article about the latest AI tool, I don't need an introduction saying this is AI and this is how it developed, because I've read this a million times in my life. So if this news article gets to me, it knows already that I know this and it will go straight to the content. So having this content completely personalized I think is super exciting. And even the example that I gave earlier about teaching, if you can then have your news reader that you like. So if you like a male voice, a female voice, or someone that looks like a dog, whatever, or a cat reading the news to you, you can do that. So we will move away from someone has created a piece of content, a video, an article that is delivered to different people to that piece of content. That article will look very different when you read it to when I read it. And this piece of video will look very different. And one world that I find super exciting in terms of Generative AI is video gaming, for example. This can again, completely customize a Generative AI can do generate new characters that only exist for you. You can then have dialogues in a video game with someone about a topic that you're interested in, instead of having a scripted dialogue with some weird characters someone has designed. So everything can be completely personalized in a world that is enabled by Generative AI.

All right, so I think we should touch on some of the risks from generative AI. So there are a lot of worries going around. What do you think are the most important concerns?

Yeah, there are so many huge concerns. That one is a lack of transparency, that there is still this problem that we don't really understand. These large language models that power generative AI, so they are basically trained on pretty much the entire world's knowledge and written content. They're large language models. Everything that has been published, every book, every piece of content in Wikipedia, the entire Internet, all of this has been sucked into it. So it understands language, but we don't really understand why. It is also so amazingly good at analyzing data. And that to me is exciting. At the same time, a little bit concerning. We don't really understand. And then there are things like we don't understand why sometimes these large language models make things up. They hallucinate. I use these tools every single day. So a good example is every magazine I write for asks for a different version of my bio. They want a 100 word biography or 200 word biography and they want to have this and that and that. And I've trained my AI on my biography or usually I say give me a draft of a 100 word biography for me. And it would create that. And every now and then doesn't happen very often, it creates one that is completely wrong, which is fascinating to me. So I recently needed to go to a conference and they wanted a biography of 200 words. I sent this over. I asked the AI tool to create one and it said I had an MBA from Warwick and I had a PhD from Oxford and places I've never been to, never had anything to do with it. And that is shocking. And you think this is so weird that this sometimes happen happens. And the interesting thing is that nobody really knows why this has happened. So it would be useful sometimes to understand how these AIs make their decisions. How they come up with data bias is another problem. That because these systems have been trained on everything that is out there on the Internet and the world unfortunately is biased and racist and everything else, it picks these things up. So in the past if you ask for examples of professions a man would do, you would get doctors and lawyers. And then if you ask for a profession for a woman, you get nurses and secretaries. And of course the world is changing and this is absolutely not what we want to see. So companies are becoming aware of this and they're making, they're coding this anti bias rule now into their AIs, which is fascinating. But Then that was recently a good example. Google Gemini has done this so effectively when their AIs can also generate images, right? So you say, I want an image of a group of Nazis, for example, if you wanted this and suddenly you have a complete diverse set of Nazis. Or you said, I want a picture of Native Americans and suddenly you have black and all different races, Native Americans. And you think that is taking it a far because now we are changing history. So there are risks in all of this. So these systems are not perfect. I have ethical concerns about this. So how we're using it, are we using it in the right way? For example, we talked about the fact that now all our teenagers have access to this AI on their Snapchat app. And there will be some teenagers that are may be a bit vulnerable that will develop a very bizarre relationship with this AI, because this AI is available when all their friends are asleep and it can answer your most intimate questions that you might have and you might develop this really weird relationship. Are we really addressing this at the moment? I don't think so. Another example that I talk about in my book is there's now an app that allows you to create your dream girlfriend, for example. And as you can imagine, most of those characteristics that you pick are physical characteristics of sizes of various body parts that you want in whatever sizes. And then the AI will create this person for you that you can then fine tune until you're completely happy that this is your dream girlfriend and they're working on a dream boyfriend version as well. And then you can have a relationship with that person. You can send text messages, you can send voice messages. Again, this person doesn't exist. And for me, he can imagine what this app is designed for. It's not just chit chat, as you can take it to a different level and ask her to send you some rude pictures of her doing whatever you can imagine. And again, for me that ethically this is not right, because I don't want my teenage boys to grow up in a world where they think they can text anyone on anything at any point in time and they were just delivered straight away. This is not a relationship between a man and a woman. So huge concerns about all of this. Also the fact that we now have so much deep fake content. There's real concern that in the within a year or two, most of what we see online could be generated by AI. And this is already happening. If you look at the latest handsets from the latest smartphones, they have the capabilities, if you take a picture somewhere, selfie of you with lots of people in the background. You say, I don't want these people in the background. And this beach doesn't look quite as nice. I take these people away. I take this hotel away, and make the water a bit bluer and the sea a bit more. The sea a bit blue and the sand a bit more yellow. And then suddenly we change everything. Nothing is real anymore. Everything we see online is somehow augmented. And that is a concern I have on a much bigger scale. There's concern around security. So AI is super powerful. It can create cyber attacks beyond anything that people can do. And it can power autonomous weapons. This is something I have serious concerns about. We now have the capabilities of having these generative AI capabilities within drones. And the US Air Force has already tested a drone swarm. So it's basically fighter jet releasing 200 drones that are really tiny, that all work with each other. And they have facial recognition technology. So they say, okay, I'm hunting down Richie here, and this is his picture. And then these 200 drones will chase you down and destroy you. And this is not science fiction. This technology is already available.

You just created my new nightmare. Now, 200 drones that know my face, we were tonight. All right, so plenty of things to worry about. I don't think we can end on that note. So for something happier, can you tell me what you're most excited about in the world of AI?

Yeah. So what I'm most excited about is that this is such a powerful technology, the most powerful technology humans have ever had access to. What I'm excited about is that this technology will accelerate so many other transformative technologies, like gene editing, like the metaverse, like 5G and quantum computing will make will enable it. So all of these different technologies will become stronger. This will mean that this technology will become even more powerful in five years time. And actually, we need those capabilities to solve some of the biggest problems we face in the world. We have a massive challenge around climate change. We have a massive challenge around inequality, be it access to healthcare, access to education, access to food. And I believe that all of these things can be addressed with the power of AI. So for me, it is giving humans the superpower to solve some of the biggest problems we're facing the world.

Potential for solving lots of really, really big problems.

I love that.

Okay, thank you very much for your time, Bernard.

Thank you so much. Was a real pleasure.