


AI’s impact on national security in the UK and US, including warnings about rogue autonomous agents, espionage risks inside AI labs, and what governments and companies are (or aren’t) doing.
No specific guests. Hosts are Katie Brescott and Danny Fortson.
Not applicable.
MI5 says AI is “supercharging” security threats, potentially enabling terrorists to develop nuclear/bioweapons. UK AI minister Kanishka Narayan says Britain will be hacked by rogue AI agents and stresses resilience. OpenAI and Anthropic allege spies inside their organizations; the FBI is investigating a Saudi national accused of planting a recording device (described as hidden in a toilet roll) in an OpenAI meeting room. Jacob Cox (former Anthropic researcher) publicly raised fears about spies.
Zuckerberg filmed drinking Guinness; Cheeky Pint podcast; Anthropic hiring a £145k PA; OpenAI claiming it solved “millennium” math problems.
Written by AI. May contain mistakes.
Explore how AI is perceived as a security threat by UK authorities.
“Like just try and talk about something else.”
Discuss the UK government's actions and concerns about AI security.
“because since we did the interview, I just saw that Berkshire Hathaway has put a further$330 million into the company.”
Examine the allegations of espionage within AI firms like OpenAI and Anthropic.
“However, in other parts of government, that's not the case.”
Analyze how public figures are shaping the narrative around AI dangers.
“Well, I will just say, maybe let's pump the brakes on the international espionage.”
Discussing the emergence of former tech employees sharing insights about AI.
“It's really interesting as just like a media, almost as a media story.”
Exploring the irony of AI taking jobs while requiring human PAs.
“Danny, did you know that my first job was working as a PA?”
Previewing the next episode featuring Luca Ferrari from Bending Spoons.
“Well, that is it for this week's episode of Times Tech.”

Get the gist, explore an idea or ask about something said.
On Times Tech, what are the key points of this episode?
Automatic transcript. May contain errors.
0:00Danny: Your heart can tell you a lot about your health. Apple Watch Series 12 measures your heart rate every five seconds with the most advanced heart rate sensing in an Apple Watch. So your vitals app, now with heart rate variability, could tell you when something is off. And your readiness score can let you know when to rest and when to push. Hear the story in every heartbeat with Apple Watch Series 12. For general wellness purposes only, not intended for clinical use or to detect, diagnose, monitor or treat medical conditions. If you felt stuck trying to lose weight, you're not alone. Enter Weight Loss by HERS.
0:35Danny: It's designed to support you in reaching your goals. And HERS now offers access to an affordable range of FDA-approved GLP-1 medications, including the WeGovi pill and the WeGovi pen. Even better, with a range of affordable GLP-1 options, HERS makes it simple to find an approach that fits your needs and your budget. If eligible, you'll get a treatment plan personalized to you and unlimited dosage changes as needed. It's weight loss designed to work with your life. Ready to reach your goals? Visit 4hers.com slash eligible to get personalized, affordable care that gets you. That's F-O-R-H-E-R-S dot com slash eligible.
1:12Danny: 4hers.com slash eligible. Weight loss by HERS is not available in all 50 states. Wegovi is the registered trademark of Novo Nordisk AS. To get started and learn more, including important safety information, Wegovi clinical study information and restrictions, visit ForHers.com. This episode of the Times Tech Podcast is sponsored by ServiceNow. Danny, one thing we keep hearing from business leaders right now is AI sounds great, but how do you actually make it work inside a company? Exactly, because most organizations aren't neat, shiny systems. They're layers of software, legacy tech, and teams all doing things slightly differently.
1:52Danny: ServiceNow sits across all that, acting as a control tower for making work move seamlessly through the organisation. Connecting people, systems, data and increasingly AI agents so that things don't happen in silos. Learn how ServiceNow puts AI to work for people at servicenow.com.
2:12Danny: Danny, hello. Hello. Can I ask you a question? Of course you can. When you go to the pub, what do you drink? Or a bar, I guess. You don't really have pubs, do you? I mean, I know what a pub is. You don't have to translate to American. But like when you're with your mates. I like Old Speckled Hen. You know, I like a good ale, especially when I'm in the UK. I was going to say, can you get that in America? Here and there, it's kind of hard to find. Usually it's like Miller Lite, you know, or something like that, which is not very good. The reason I ask is there's been a lot of hilarity over here in the UK, as you can imagine, with Mark Zuckerberg being filmed drinking a Guinness.
2:51Danny: So swilling it around his mouth rather awkwardly. Looks like he's chewing it. Yeah. Pulling a bit of a face at the bitter taste. Yeah. So there's this, the Collison brothers, the guys who founded Stripe, they have their podcast called Cheeky Pint. And they've mocked up like a pub set up, I think, in their headquarters. And then they have all the great and the good come through, like, you know, Elon Musk. And I don't know if Bill Gates has been there, but like. And they are Irish, right? And this is also important. Yes, they are Irish. Correct. Correct. And then the whole, the pot is, you know, they have a cheeky pint with a tech leader, a billionaire, whomever, and talking about what's happening.
3:28And Zuckerberg was on there with Alexander Wang, who's the founder of Scale, who then he brought in to kind of overhaul their whole AI effort. And it was funny because they're talking to Alexander Wang. And then in the background, you see Mark Zuckerberg like two hand his Guinness. Guinness. And he's just like, and then he has a swig. And then he's like, like, like, he looks like a cow chewing its cud, you know, where you're like, what is what are you doing?
3:54Danny: My question to you then is, is this because Guinness is not much drunk? Where you're from? Is he unfamiliar with it? Or is this a Mark Zuckerberg weirdness? I think it's both, you know, because he does brew his own beer, apparently on his compound in Hawaii. I went to the Guinness factory in Dublin. Did you chew it? I didn't. I mean, I'm not a natural Guinness drinker. But I know how to drink a pint in a pub without looking like a weirdo. We all have our skills. We all have our skills. Oh, goodness. Well, anyway, should we get to the pod? Let's do it. Hello and welcome to Times Tech, where we unpack how technology is reshaping our everyday lives.
4:36Danny: I'm Katie Brescott. And I am Danny Fortson. And this week we are talking about, can you believe we're talking about AI this week? I think we should have a ban one week. Like just try and talk about something else. Anything else. But we're talking about how AI is impacting national security because a few things have happened. Oh, it's been such a hot topic over here. It feels like all of the red lights are flashing over it. So this week, MI5 has warned that AI is supercharging security threats to the UK. Meanwhile, Britain's AI minister, Kanishka Narayan, has said it's inevitable. Britain's going to be hacked by rogue AI agents.
5:13Danny: And plus, there's this extraordinary news of a hidden recording device inside OpenAI. Yeah, sophisticated recording device that was, quote unquote, planted inside the premises. And they said they have notified the FBI who are investigating. And this all, of course, comes at a very interesting time with questions around security and the arms race with China. And in the last few days, Jacob Cox, remember him, the former anthropic researcher who left, tweeted about his worries for the future of the human race. And he's raised alarms now about the possibility of spies inside both anthropic and open AI.
5:55Danny: Yes, the man from anthropic who said that AI could kill us all. But just when you thought AI was going to wipe us all out and take all our jobs. Never fear. There is heartening news, as reported by the Times this week from Anthropic. Anthropic is hiring a PA, a personal assistant, for a salary of£145 ,000 a year. And they're going to do jobs like managing calendars and monitoring inboxes. So all of the things that my instinct agent does. Maybe we still need people after all. Yeah, there's hope for us yet. And in case you haven't heard, Katie, we're doing two episodes every week these days. On Fridays, this one runs, of course, the big news stories of the week.
6:37And then on Sundays, we have our interview episodes with the movers and shakers of the tech world.
6:44Danny: And this week, we've got a really, really good one for you. I've been speaking to the founder and CEO of the company with, I think, the best name in tech, Bending Spoons. They're an Italian-based tech firm that buy up old tech companies like AOL, remember them? Evernote and Vimeo. And I asked the founder, Luca Ferrari, the question on everyone's mind, sort of, why are you doing this? But it's clearly an idea that's really been capturing the imagination because since we did the interview, I just saw that Berkshire Hathaway has put a further$330 million into the company. So that's Warren Buffett's investment company.
7:20Danny: Can we say that now he's retired. I don't know. But he that they now they now own 4 % of bending spoons. Anyway, so really, really interesting interview with Luca Ferrari. But let's get to today's big stories, starting with the one closer to home for you, Katie. Yeah, I mean, it has felt like everybody here has been sort of waving their hands in the air about AI and security this week. So starting on Monday with our AI minister, Kanishka Narayan, who I should say is incredibly popular with the tech community here. He used to work in venture. And, you know, if you go to any of the big tech events, he's always there.
8:00Danny: He's very, very plugged into the ecosystem. And even though the government scrapped the Department for Tech, they kept him and he's now in the cabinet. So it's kind of his first big intervention in this new government that we have. And he basically said, Britain's going to be hacked it's inevitable by autonomous rogue ai agents and he talked a lot about the importance of resilience and he said we just you know need to do more to protect ourselves and then hot hot off the heels from him saying that we had the head of mi5 our home security service saying ai is supercharging security threats and that includes the potential development of nuclear of bioweapons by terrorists.
8:40Fine.
8:40Danny: Cool. Nobody panic. Yeah. The MI5 director general said, which I thought this quote was interesting, he said, human history has given us no experience of facing an actor which is smarter than we are, is capable of outthinking us. This is genuinely novel and deserves the attention it is getting, including in MI5. Just, you know, again, nothing to see here. So what's happening? Because I feel like every week we get, you know, whether it's Jacob Coxon or somebody else be like, guys, we're screwed. And there's like, let me wave my hands around. Let me freak everybody out. Is there anything like happening?
9:17Is the government doing anything? Or is this kind of a sense of like, don't worry behind closed doors inside MI5, we've got a crack team on this, we're doing this, we're doing that. Like, is there a sense of like, action or anything like that?
9:30Danny: Yes. So I think that behind closed doors, there is a sense of action and there is a sense of cooperation with their U.S. security counterparts. And I don't think that's gone away. And I've heard it's interesting at tech conferences at the moment, you do hear representatives from the security agency speaking. And I haven't seen that actually so much before. So I think there's a genuine concern about it, a genuine desire to get the message out that they are working on this. However, in other parts of government, that's not the case. And actually, I'm getting a sense of lots of anxiety about what on earth is being done about AI.
10:12Danny: It wasn't mentioned, for example, very much in Andy Burnham, the new prime minister's speech at the recent Labour Party conference. You know, the message and the concern from government has really been around inequality, revving up the economy, what on earth they're going to do in the next budget. And actually, this feels to me that it is being left to security services, as I say, but also the AI minister, who clearly does really grasp the nettle on it and see some of the concerns. But there are so many fish to fry at the moment. There are so many other concerns for the government that it feels just like another piece of worry.
10:54Danny: We had news from the AI minister this week that AC, the security institute here that looks at AI bodies, is now going to get first sight of models coming out of companies from the US. That had fallen by the wayside, so that's come back. So there is action, but there is certainly also anxiety that perhaps the government isn't as a whole taking it as seriously as maybe it should be. Right. And then you have in the States where, you know, the frontier labs are going to the White House and be like, guys, guys, guys, don't worry. We're going to police ourselves, which is just to your point around what's actually happening amid all of the freak out is it doesn't feel like much is happening.
11:37Danny: You probably hear this view all the time, but I spoke to a US contact recently in one of the big tech companies and they were saying it's all very well to mock Trump for this knocking the heads together of the leaders of the AI companies and just trying to get them to sign an agreement, which doesn't really mean anything. But actually, the tech is moving so quickly that it's quite a sensible approach in some ways, getting them to talk, getting them to appear like they're doing something, because regulation is just not going to keep up with that. And that takes us neatly on to the other big security story of the week.
12:12Danny: And Anthropic and OpenAI have said that they're the target of spies. And lo and behold, the FBI released details of this extraordinary story about a Saudi Arabian national who'd been working for OpenAI for a year, who allegedly planted a bug in a meeting room in a toilet roll. Why do you have a toilet roll in a meeting room? I wondered that too. I was like, surely that's the giveaway. Apparently it was found a couple of days later. And I should say, you know, this is all allegations, but they've released this footage of this engineer sort of peering behind a screen in a room that he wasn't supposed to be in to look at this bug.
12:52Danny: and they found apparently details on his computer of him researching recording equipment. But that shows just how this has become an enormous national security issue that there are potentially spies within this company finding things. Go on. Well, I will just say, maybe let's pump the brakes on the international espionage. It's a great story, isn't it? It's an amazing story, but it also might be an idiot. who's like, you know what? I want to be a spy. I've got it. I'm going to hide a recording device in a toilet paper roll in a meeting room. No one will find me. I'm glad you're not representing him in court.
13:36I mean, come on. Yeah. And so investigations are ongoing. But I do think, you know, we're in such a weird time. It's like, you know, this guy, Jacob Coxon, who's 27 years old and left Anthropic. He was at OpenAI, then he went to Anthropic, and then he left famously and was like, uh-oh, I'm really freaked out. This may kill us all. And it's amazing. He's gone from like just a rando, random guy in Silicon Valley to suddenly like a respected authority on what is actually happening inside these companies, the danger this poses to humanity. It's really interesting as just like a media, almost as a media story.
14:17it's like there's this 20 something person who's been working has a working life of three or four years sees something worrying leaves post something on a twitter account that has zero followers all of a sudden he has whatever hundreds of thousands he's on you know the daily show on cnn on every you know major network and now he's talking about how there's definitely spies in there and everybody's like yep yep yep yep and i'm i mean i'm sure he's i would not be surprised if he's right? But it's interesting how this kind of random Silicon Valley kind of worker bee has become like this global authority on what's happening inside AI and the dangers it poses.
14:59Danny: But isn't this because we don't hear that many of them? And isn't that also because they've all signed tons of NDAs, I imagine? Knowing friends who've come out of tech firms and are worried about what they can say? Isn't it quite unusual to have someone who's been in the thick of it? We're all desperate to know what it's like in Anthropic and OpenAI. I don't think it's rare at all, which is why I find it so weird. Because it feels like there's a whole oeuvre now of like, hey, I worked at OpenAI. I worked at, today I'm leaving Anthropic. I'm going to go write poetry because... But do they do the radio and TV rounds?
15:34No, but that's what I'm saying. He is. Yeah. And he's leaning into it. And like, people are lapping it up. I just think it's really interesting. It's not like he's like, hey, I was the former head of this. I was the former executive doing that. It's like, I'm just a dude who was at this company for six weeks. And he's like, let me tell you what's really going on.
15:54Danny: But I don't think we get many of those dudes speaking to us properly. I mean, look, if you've worked at Anthropik or OpenAI and you want to come on the pod and tell us what it's like, I'd love to hear it. All of it. So would I. Yeah. After the break, we're going to talk about Anthropik's new role, exciting, new, modern, PA roll for 145 grand a year. Pounds or real money? Pounds. How dare you?
16:24Today's episode of the Times Tech Podcast is sponsored by ServiceNow.
16:28Danny: ServiceNow describes itself as an AI control tower for business reinvention, connecting people, systems, data, and AI, so workflows more smoothly across an organization. The platform integrates with different clouds, models, and data sources, bringing them together in one place to help your team. To learn more about how ServiceNow puts AI to work for people, visit servicenow.com.
16:55Danny: Your heart can tell you a lot about your health. Apple Watch Series 12 measures your heart rate every five seconds with the most advanced heart rate sensing in an Apple Watch. So your vitals app, now with heart rate variability, could tell you when something is off. And your readiness score can let you know when to rest and when to push. Hear the story in every heartbeat with Apple Watch Series 12. For general wellness purposes only, not intended for clinical use or to detect, diagnose, monitor or treat medical conditions. Requires iPhone 11 or later. Hey, I'm running out to drop off a stack of letters and packages.
17:28Danny: Need anything shipped while I'm gone? Save the gas. Stamps.com lets you print postage and schedule pickups right from your desk. Sync your contacts into the address book. Get detailed accounting insights and get notified the moment your certified mail is delivered. All from my desk? All from your desk. Join the 4 million businesses using Stamps.com right now with code podcast and get your free welcome gift. Stamps.com code podcast. Taxes and fees apply.
18:00Danny: Danny, did you know that my first job was working as a PA? For, what's that dude's name? That's right. Peter Jones from Dragon's Den. Exactly. How was that? It was the most amazing experience, actually, because I got to work across his portfolio of startup businesses. I was working on a TV show. He started a TV company. We started an academy with Gordon Brown's government. So we did all those sorts of things. And my job as his PA was to manage his diary, to make sure he knew what he was doing when he turned up to meetings, and generally sort of be like a second brain, I'd say. Did you make 145 ,000 pounds?
18:41Danny: I didn't. I think my starting salary was a fraction of that. Certainly wasn't six figures. Well, you know, that was only, what, two, three years ago, you know. I mean, yeah, obviously. You're still in your 20s, obviously. 26. Yeah, so. um but anyway anthropic is now advertising for a pa for 145k and this sent our business desk at the times into spasms of laughter and everyone was talking because it was like aha these guys and they're trying to replace us all and they keep talking about how all of our jobs are going to go out the window and they're going to do all our admin for us and actually they need a pa but But isn't that kind of like, not that surprising?
19:28Because I just like the further we get into this AI boom, I'm just like this notion that we're all going to lose our jobs, I think is so divorced from the real world.
19:40Danny: But they keep saying it. I know, but it's so like divorced from the real world and just increasingly like just lacks credibility in any way, shape or form. But that's why this is so funny because they keep going on like, oh, you're all going to lose your jobs and we're going to be on universal basic income. And what did Dario Amadei say about AI? That quote you always throw out about geniuses in a data center? Yeah, country of geniuses in a data center. And actually, they need someone to do their admin who must have, and I quote, immaculate systems to ensure no details stick through the cracks and deliver error-free execution.
20:17Did you deliver error-free execution?
20:19Danny: No, I didn't, but... No, absolutely not. What I did deliver was personality and a very high EQ, I would say. Because when you're working very closely with someone who is under an inordinate amount of pressure because they're juggling lots of different businesses, plus they're a public persona. So they're doing lots of events and they've just got to make lots of decisions. You need to have a real understanding, which I don't think AI could ever have of. If you're managing their diary, you can't just fill it with meetings. You need to think that meeting will need a half hour chat with so-and-so beforehand.
20:59Danny: He will need to sit for an hour after that meeting and silence because it'll be so horrendous. You know, there needs to be thinking time booked into that. Or I know it's his daughter's sports day on Thursday. And I just don't know that AI can ever quite capture that closeness of working with someone. Maybe, but, you know, or just you look at someone, you go, God, you're tired today. Let's ditch the, you know, yeah. Yeah. Well, I think it's interesting that, you know, this popped up and it created great hilarity on the desk in the same week or in the same, within a few days of OpenAI putting out like, Like, hey, we just solved these millennium mathematical problems that were created at the turn of the century.
21:48And they were like the greatest, most difficult, most complex mathematical puzzles the world has ever seen. And they were all just solved by OpenAI. And it was really interesting. I was listening to, not to pump other people's pods, but I was listening to The Daily and they had on a mathematician. And he was like really emotional. about it because it got to like what is your purpose if there's this new thing that can just like has the intellectual horsepower that just like no human can ever muster so it can just solve things in a way that just humans can't like what's left for you and it was really interesting interview because i was like i don't really care about math or certainly not like these very very deep complex problems but it was just like this like this kind of look into which is why i think if you're inside anthropic or inside open ai back look what we can do there's going to be nothing left for humans and in these pockets you know you're making incredible breakthroughs but then you're like to then back out and be like you know what there's not going to be anything left for any of us we're all going to lose our jobs anything that you know you can do with a computer ai can do it better all that stuff I just don't think it kind of again I think it's completely detached
23:09Danny: from the reality of inertia of organizations I mean I think about the work I did as a PA a lot of it was knowing everybody in his life and I hate the word ecosystem but ecosystem and those dynamics and you just I don't believe you can program that and I think about our office manager here at the times jackie and the way she knows everybody on the desk and their personal situations and how carefully she thinks about that when she's doing rotas and the speaking to people and just these jobs are far more than you know what they say can be replaced frankly and also they by the way their systems aren't good enough ha ha ha that's what we thought anyway to do, you know, as they say, immaculate systems to ensure no details slip through the cracks.
24:01Danny: Maybe you do need a really good person at the helm. Yeah, none of my systems are immaculate. I'm a mess. As you've often commented on the number of tabs I have open at any one time. That's shocking. Yeah. Right now I have probably 25 tabs open. But anyway, that's just, you know, beautiful mind. Don't judge. This is a safe space. It's the creative brain at work. Exactly. Well, that is it for this week's episode of Times Tech. We'll be back on Sunday with that fantastic interview with Luca Ferrari, the chief executive and co-founder of Bending Spoons. Here's a clip. One of my co-founders, Matteo, is a big fan of the movie The Matrix.
24:42And there is a little kid at some point bending a spoon with his mind through telekinesis, presumably. And so, Matteo, we're brainstorming names, brings up the name. And we liked it a lot for a few reasons. One, it just captures the idea that the mind is powerful. It can do a lot more, perhaps, than you intuitively think it can. And then that power is unlocked through plenty of practice, perseverance, hard work. And those are things, call them values or ideas that resonate with us. And also, we thought the name was kind of a little bit funny. We like the idea of not taking ourselves too seriously.
25:19Thank you for listening. And if you don't already, please do follow us on YouTube or wherever you get your podcasts so you never miss an episode.
25:28Danny: See you next week. Bye-bye.
25:49Danny: could tell you when something is off. And your readiness score can let you know when to rest and when to push. Hear the story in every heartbeat with Apple Watch Series 12. For general wellness purposes only, not intended for clinical use or to detect, diagnose, monitor or treat medical conditions. Requires iPhone 11 or later. When it comes to managing money, forget the hype and look at results. Bill has a trillion dollars of secure payments powering our BillPay tools. Instead of just moving money, Bill is powering the financial operations of nearly half a million customers. So stop the guesswork and start scaling with the proven choice.
26:25Ready to talk with an expert?
26:26Danny: Visit bill.com slash proven to get started and grab a$150 gift card as a thank you. Terms and conditions apply. See offer page for details. Still waiting in line? Again? That's time you'll never get back. Save time and money with Stamps.com. Over 4 million businesses have skipped the line with Stamps.com. Join them to save up to 90 % off carrier rates from your computer or phone right now. Print postage for certified mail, registered mail, and packages in seconds. Then schedule a pickup right from your home or office. For a limited time, go to Stamps.com and use code PODCAST for a free welcome gift.
27:03Danny: Taxes and fees apply.
MI5 has warned that AI is "supercharging" security threats to the UK, while Britain’s AI minister said it was “inevitable” that Britain is going to be hacked by rogue AI agents.
Danny and Katie look at the growing security fears around AI, including the FBI investigating a case involving a former OpenAI employee accused of planting a covert recording device at the company’s San Francisco office.
This comes days after Jacob Coxon, the former Anthropic employee who warned about AI’s existential threat, raised alarms about the possibility of spies inside Anthropic and OpenAI. So, what does this tell us about the AI industry and the potential dangers it poses?
Plus, after years of predictions that AI will replace white-collar workers, Anthropic is reportedly hiring a human personal assistant for a salary of £145,000.
Read more: Chatbot developer Anthropic seeks PA — but only humans need apply
Get in touch: [email protected]
Presenters:
Katie Prescott, Technology Business Editor, The Times
Danny Fortson, US West Coast Correspondent, The Sunday Times
Producer: Marnie Duke
Executive Producer: Priyanka Deladia
Image: Getty
Hosted on Acast. See acast.com/privacy for more information.