AI Could Destroy Humanity... Should Investors Be Worried?

AJ Bell Money & Markets

The episode explores the dual nature of artificial intelligence, weighing existential risks against its practical utility as a research tool

Key takeaways

• Anthropic's leadership warns that AI advancement could cause massive internet damage within 12 months.
• NVIDIA CEO Jensen Huang advocates for industry self-regulation rather than heavy government intervention.

Main topics

• AI existential risks and regulation
• AI as an investment research tool

Notable quotes

Use it as a lens. It's not a crystal ball.

Conclusion

While technological and economic shifts present significant risks, pro

Transcript preview

Speaker 5 (0:09) Hi, welcome to the Age of Bell Money Markets podcast. I'm Dan Coatsworth and joining me today is Sarah Coles. Hi there, Sarah. Speaker 2 (0:15) Hi, Dan. It's good to be back. So this is actually my first podcast after the summer break and it does feel like I'm coming back to quite a lot because there's just so much to talk about, including, rather alarmingly, the end of the world. Speaker 5 (0:27) I know. Tech experts are now talking about whether AI can kill off humanity. So with that in mind, we're going to chat a lot about AI in the show. I also spoke to... tech journalist Will Guyot about whether AI can actually play a positive role for investors. We're also going to explore what else has been moving on the markets, including headline grabbing bond sell-off. Speaker 2 (0:48) Yeah, and on the back of your really interesting chat with Danny last time about student finances, we actually are going to hear from a couple of listeners about the approaches that they've taken. And obviously, it's something that I know I'm facing this year, and it's weighing quite heavily on my mind at the moment. Plus, I've got some less cheery news about what the new term can mean for relationships and how to protect your finances from the fallout of it. Speaker 5 (1:09) There's been a lot of interesting data out this week, including wage data. It gives us an idea about how the state pension will rise next April. So we're going to have a look at the implications of that. There's also some savings data out. So we'll have a look at the world of ISAs with that. And of course, we've got inflation, just as we're recording this, revealed that it's gone up by 3.1 % over the last month. Speaker 2 (1:29) I have to say, I think now my... teenage son has returned. My spending's definitely gone up more than 3.1%, but that's just me. And finally, there's a new report into the gender pensions gap. So we'll explore the findings and share some tips of what people can do to boost their pension if they're away from work or working part-time. So there's loads to discuss, but we really should start with the end of the world. It does feel like we should avoid having the elephant in the room for too long. So senior tech bros have been warning that AI can be getting too powerful too quickly and could wipe us all out. So Has this had an impact on the sector for investors? Speaker 5 (2:03) I mean, it's like, it's pretty alarming stuff. So this is sort of, I'm sure most people listening or watching this podcast will have sort of caught this news. But if not, we had the boss of Anthropic, which is one of the AI platforms that owns, Claude, coming out and basically urging all the AI sort of big platforms to voluntarily work together, try and set some standards. And what it was, the crux of this sort of thing was he's warning that AI technology is actually advancing dramatically faster since the summer and could be capable of taking over the entire internet and causing hundreds of billions of dollars in damage in the next six to 12 months. So that's about as alarming as you can get. And we've got a chap called Jacob. Coxon, who's a researcher at Anthropic, had said last week that he'd actually resigned from the company because he said the people building AI earnestly believe it could kill us all by the end of the decade. So how on earth do we deal with all that information? So, of course, this is naturally led to people debating about should there be a kill switch if everyone's now thinking that, you know, what if these AI platforms get out of hand and cause terrible damage? Donald Trump has come out and said, you know, of course we don't need all this stuff. You know, just let them do their thing. NVIDIA boss, Jensen Huang, he's come out and dismissed the idea of really tough AI regulation. And he was talking at a conference in San Francisco saying, you know, companies should be responsible for the safety of their own tools and they should decide themselves whether they should, you know, should they release or not release dangerous products. So, you know, I think that obviously when you've got, the boss of the person who, you know, as the company's at the center of supplying chips into the AI industry, they're coming out and sort of being quite relaxed. The president of the United States being quite relaxed. He's like, who's right or wrong here? Now, Jensen Huang did say, he said, run as fast as you can, but if you feel that if you're not in control, take a pause. Obviously, that's sort of, you know, that's a sort of a sensible approach. But Sarah, I wanted to just... I wanted to sort of mention this on the podcast, and I don't know whether anyone obviously watching it has cottoned on to it. There's another company called TSMC, which is Taiwan Semiconductors. This is the world's largest AI chip manufacturer. Someone there came out and said it was basically comparing AI to a three-year-old Superman, basically saying AI can be quite powerful, but doesn't know what kind of damage it can inflict. And I thought... I thought that was a pretty accurate description. And it's like, were you alarmed by all this news when you read it? I Speaker 2 (4:46) think, yes, obviously. It's so alarming, you just kind of, your brain protects you from it by just assuming it can't be right. I mean, I think it is difficult. Obviously, you know, lots of people saying, no, it's fine, look the other way. It's obviously always a bit of a red herring. Sorry, it's always a bit of an alarm bell. when people kind of try and tell you not to worry. But I mean, in terms of how the markets have responded, have the markets kind of taken it with a pinch of salt or are they kind of thinking this could mean more controls on AI and therefore it's less of an investment opportunity? Well, I Speaker 5 (5:16) think we've had, I don't think anyone could disagree that this AI has been the hot investment theme for the last couple of years. I think people think it's all about this race to who has the most powerful, the most sophisticated system, how are businesses and consumers going to be using it? It's all very exciting. This has been a really important wake-up call that there are major risks involved with AI. And therefore, any company connected to this world needs to be aware of it. Any investor thinking about opportunities here needs to understand that some things could go wrong as well as go right. Now, the headlines that we've seen would suggest if something goes wrong, it could go very, very wrong. But the idea that people should be reappraising AI investments, I think this should be a positive. It should remind people that you've got to think about what could go wrong as well as go right. Now, obviously, it puts a spotlight on will there be greater regulation? What happens if these things are getting out of hand? Are these companies just developing things too fast? Does this mean the pace of... earnings growth for the broad AI sector could actually be a little bit slower. So I think this is a natural sort of discovery moment here. Markets wobbled a little bit for this for a couple of days. It didn't help that they also had worries about a high oil price, high inflation, what that meant for interest rates and stuff. But I think a sort of a more considered rational approach is probably quite a good thing, taking some of the heat out of the market. Speaker 5 (6:58) Yeah, or perhaps we could see in a couple of weeks is, you know, all these companies come out and say, like, yeah, we'll take the foot off the pedal a little bit. But I guess it's a natural next step to try and think about these things. Speaker 2 (7:13) And I guess presumably it's also a big boost for, you know, all those companies that could benefit from masses of prep. So lots of, you know, tin can companies and toilet roll companies and that sort of thing. They can have little canics and build their own bunkers. Speaker 5 (7:27) It's true because we have seen, you know, the last few days we have seen lots of people sort of looking at healthcare stocks and utility stocks and going, oh, crikey, perhaps I need some boring stuff as well as sort of the excitement in their portfolio. Speaker 2 (7:39) I mean, I guess AI is always going to bring a little bit of excitement and we've obviously had these scare stories for a long time. But actually, you know, you've also been exploring some of these. positives in the way that AI can actually help investors. So Dan recently caught up with a tech expert who lives and breathes AI. So let's hear that interview now. Speaker 5 (7:59) A lot of people get into grips with AI and seeing how useful it can be. And that actually includes asking for help with investing. So here to talk about AI and how people perhaps are using it from a financial perspective is leading tech journalist Will Guyot. So Will's a regular commentator on BBC Radio and LBC. We're very excited to have him on the Age of Bold podcast. Speaker 4 (8:20) So hi there, Will. How's it going? Yeah, very good. And I'm here for the reason to talk about AI, but hopefully pick up a few tips from you at the same time, because I can talk quite confidently about AI, but investing is something that's quite new to me. And actually, I have to say, AI has helped me to an extent. So, yeah. I mean, this is, it's a Speaker 5 (8:39) fascinating topic, because I think lots of people say that they, you know, investing is not always straightforward. People often... They go to Google and, you know, get some help. But obviously AI, the way that it serves up the information is a lot, I guess it's a lot easier to understand. So, I mean, from when you were doing a bit of research then to sort of think how can AI actually help me, what did you find? Speaker 4 (9:01) Well, I know having written in my other parts of my career, 200 page annual reports for companies, just how hard it is to actually find the information about how companies are performing. So if you run something like an annual report through an AI chatbot and say, pick out all the pertinent things to me, I'm considering investing. It's really useful. Never trust anything AI tells you at the moment, though. That's the one thing I'm going to say. Use it as a lens. It's not a crystal ball. It's not going to. predict the future. And it's not always accurate because until very recently, AI convincingly told everybody that I was hosting the weekend breakfast show on LBC, which would have been a very big surprise to my boss at the station because it isn't something that I was doing, but it said it with incredible conviction. So especially if you're looking at something like investing, take this as information, maybe look elsewhere, see what other people are saying about these bits and pieces. It's quite useful as a research assistant. at the moment, I'd say, or a bit of a sparring partner. But don't take it as the ultimate advice of, you know, and you're the person who's the expert on this, but I wouldn't use artificial intelligence at the moment as my financial advisor. I might use it to say, I've looked at AI and take it to a human being and say, what do you think about these two or three things? That's probably what I would do if I was using it at the moment. But importantly, it's moving really quickly. And I would say today is... you know, infinitely more better and more insightful and getting closer to accuracy than it was even a couple of years ago. I Speaker 5 (10:37) think I just wanted to ask you whether you thought that, you know, when you're using AI, whether it's for investing or whether it's for anything, does it sort of encourage you to take shortcuts? And ultimately, could that actually lead to mistakes? Because you're not really putting as much effort into your research because it's trying to make your life easier. Speaker 4 (10:54) Oh, yeah, exactly. And there's so many reasons in that. outside of investing in everything else that AI makes you dumber because you literally put something into it many people just take that and accept it I would say you know if I'm looking at something like this trust it in the way you treat a really well-read friend the bloke at the pub who knows a lot about something not the person sat at the end of the bar coming up with you know different theories for different things so I would sort of say It's kind of, it's a support at the moment. You know, it's kind of right on certain concepts and bits and pieces of information, but it can absolutely still get stuff wrong, but give you false answers with such conviction that you think it's true. So, you know, that's kind of where I see it at the moment. It's useful. It's getting more reliable, but don't make it your last line of defence. Yeah. Speaker 5 (11:46) Yeah, because I think it's Speaker 4 (11:47) natural to sort of type in. to ask Speaker 5 (11:50) a question like, is this a good investment? And I wonder if there is, like you say, a big risk that people actually think that this is fact. It's not just opinion. Speaker 4 (12:00) Yeah, and that's the big challenge of AI at the moment. These companies have designed these large language model chatbots to agree with you most of the time, so you spend time using it. If it said... Dan, I think you're a clown. You've got this completely wrong. You're going to go, well, stuff this and go somewhere else. But if it says, oh, great suggestion, Dan, you know, you're on the right track, et cetera, et cetera, et cetera. It has a habit of sort of talking authoritatively about something and leading you to the courage of your own convictions. And not necessarily, again, in the world of investing, but I hear that countless times in other places. We've heard of people who've, you know, people who may have had other issues. mental issues and other things outside suggesting stuff to ai and it's saying that's a great idea and literally selling their house to go and create an idea which was absolute guff which the ai chatbot had agreed with because it wanted to be seen as siding with um the human that was operating it so there's always an element of that and like i say if we go back to the primary point there's a lot of use in this as a research partner but don't don't let it think for you Speaker 5 (13:06) Yeah, I mean, I think there's sort of two ways we can look at AI from an investment perspective. One is the general public just asking questions and asking for more information. The other one is kind of like businesses, how do they use AI? And so we've obviously got lots of fund managers who are picking sort of investments on behalf of. of investors. I wondered whether you'd come across any sort of interesting information about how might fund managers or anyone in the investment industry actually be using AI internally? Is it simply to just run spreadsheets, do calculations, or is there any sort of evidence they're now actually doing all the work that they used to do themselves? Speaker 4 (13:47) I think that a lot of them are still at pilot stage. And I know there was a survey from Mercer's of 131 asset managers in this year. And I think 55 % said they've integrated it into at least one investment process, but most of them are still at pilot stage. And, you know, nearly 20 % of these asset managers are not even close to kind of even trying and using it at the moment. And it's mostly concentrated at the moment on the research and the idea generation stuff. It's not touching the portfolio. folio or the execution of trades yet that's not something that they're letting chatbots um yet do but i would imagine um and i'm not super close to this but i know in every other line of work in industry that i come close to there are groups of people thinking