the one thing that kept jumping into my head was my laptop now has been infected by copilot. I to go and install it every second week after an install. My phone AI is starting to creep into the interfaces.
I obviously use VPNs everywhere because of of my profession, cybersecurity. I I use it to hook into infrastructure and work. I I use it when I'm traveling across airports, etc. But if the technologies are starting to become invasive on my devices and of themselves, and as these privacy concerns lead to more I would call loosening of restrictions around what's available and what can be curtailed and what can be consumed, is the and I and I don't like to use the word nanny state, but essentially
that all knowing, all scene surveillance, are we giving away a huge chunk of privacy without even realizing it every day? And it's feeding into a trend I'm seeing where millennials and and and Gen Z, etc., are going back to the Nokia thirty two ten, the brick phones, etc., because of notification fatigue. But I'm really at the point where I'm gonna bin it all because of privacy concerns because everything seems to be sucking a chunk of my my my digital life out out through the side door. What what's your what's your
take on on on on what's happening there as well.
paul (35:45.316)
We are. So yeah, there's there's kind of two sides to it. The one that I went through in great detail is what's happening from the governments and tech companies, right? It's not these are working hand in hand. The tech companies have everything to gain by implementing these identity verification schemes because they get to learn more about everybody that they verify. So even with children, for example, Facebook will know the age category of every child. so that's on one side. The other side is absolutely.
We are inviting it in, but it's it's worse than what you might believe, right? So I've been I I've said for a very long time that adding AI to email or any software application or system that has the ability to find and record and store sensitive information or to take action that could cause harm is a very bad idea.
And but I'm not going to die on that hill. I've written about it, and if somebody wants to accept responsibility for themselves, all and well and good, you go ahead and do that. But here's where I've got a problem. ChatGPT, so OpenAI has released a Chat GPT plugin that makes it easy for Apple users.
To add that plugin to Apple Messages on their Mac computer. And the reason that's very problematic and one of the worst things I've seen in 30 years is because most people will think of the messages app on your phone as text messaging, but most people don't use SMS. But a lot of people use Apple iMessage. So the Apple Messages app has three services built into it.
SMS, RCS, and iMessage. RCS is the message service that you that an Apple user would use when they're texting somebody with an Android phone. It just happens automatically. iMessage is one of the few communication channels that has true end to end encryption. And I met an investigative journalist from a national broadcast broadcaster in the last week to talk about chat control.
paul (38:10.711)
Which is a EU spyware surveillance law. And they told me that they relied on iMessage quite heavily when they spent a lot of time in China for privacy and personal safety. Because the FBI has even stated that they can't access any message content between two people who use iMessage to communicate, even when they have a court order. They've been trying to pressure Apple for years to put in place some kind of so-called
Safe legal backdoor so they can have access. And Apple has always refused because it's end-to-end encrypted. Even if you use Siri to read, write, or send an SMS or an iMessage, it still can't be requested by authorities because Apple doesn't have access to it. So it's secure. Now, encryption isn't broken. However, if the entire purpose of having end-to-end encryption is to make sure that no third party can access that.
conversation, then this plugin has destroyed the entire benefit of the end-to-end encryption. Since I was interviewed by Fortune and did a technical analysis, some security professionals, which beggars belief, has come back to say, but it's no different to taking a screenshot on your phone. And I'm scratching my head, but not surprised, because every industry has people who probably should change their work.
My response to that is yes, but if I text you and you and I are having a private conversation, I'm trusting that you will not take a screenshot and fax it off to the FBI. If you do, it was your decision and you broke my trust, and that's not cool. You know, most people are not going to do that. I'm sure people have done it over the years. However, if I add that plugin to my
Apple messages on my computer. That means OpenAI technically has the ability to find and record and store and analyze every conversation I've had with every single person for the last five years. Because technically, when I set it up and I give it certain permissions, it can access every conversation. Because a lot of people don't realize this, but when you swipe
paul (40:38.691)
To delete a conversation with somebody. As soon as they send you a text message the next week or next year, that entire conversation will come back if it's stored in the iCloud. So it's not just because it's deleted on your phone, it doesn't mean it's the conversation is gone. So OpenAI is saying that this is to give people the convenience of being able to schedule messages and then do a search across all your conversations and find that recipe or that person, that conversation, that phone number, that email address.
That means the US government can do exactly the same thing across the entire world of conversations. Because if I send another iPhone customer a message and I'm not in their contact address book, it will say maybe Paul because it has my details in in the cloud. And now what you've got is a social graph of everyone. And if I add it and I don't tell you.
You will never know. So every person in the world now could should reconsider using Apple messages. Now most people will use and it'll be fine, but don't use it for anything that should should be very private. Because anybody you're communicating with might have added open AI to that. And that's not the same as adding it to email or your desktop or a browser, which obviously has its own
Bad concerns and dangers and probably so you shouldn't do it. But I'm going to die on the hill of end-to-end encryption because it is implied that that will not actually be secure. Proton came out and cited my analysis and agreed with me. The CEO of Signal has already said also, as soon as you add an AI agent or client side scanning, you effectively have a backdoor. Now don't know if she used those words precisely, but
It removes the benefit of having end-to-end encryption. And iMessage is in a category of very few services that include iMessage, Signal, Session, Proton. Now you also got the likes of Facebook, Messenger, WhatsApp, and Telegram, but there's reasons as to why the highest risk individuals wouldn't use those services. So Apple would not have foreseen this either because there are software applications that you install on a Mac.
paul (43:07.735)
That require the permissions, right? The hard disk, the contacts. But you're installing software for something that requires those permissions. And I don't believe Apple would have seen foreseen this situation where a company, because Claude Anthropic has built something similar. Anybody could build an application that does this, but OpenAI has gone out of its way to build a consumer product for millions of people, specifically.
For a messaging app that has end-to-end encryption. For me, that is one of the most irresponsible things anybody could do. My team built the world's very first security chatbots for every messaging platform. They were amongst the first chatbots in the world. From 2015, the first chatbots for Slack, HipChat, Skype, Messenger, and Telegram. I was given talks at developer conferences.
Trying to encourage developers to build these chatbots like ChatGPT with ethical design in the foundation and thinking about the future. So I have designed a chatbot plugin for messaging services, right? So the one for Messenger was used by Al Jazeera. You just post a link and it told you if it belonged to a website or social media account that was classified as disinformation, far left, far right, satirical or mainstream.
That it was a the stupidest, dumbest bot you'd ever come across. That's the only thing it could do. So I I know the tech stack and I've built bots like conceptually the chat GPT for Apple messages, but I would not under any circumstances build a service that has the ability to read and store information, even if I had the best security in the world, the best encryption, the best of everything, the best intention.
I still wouldn't do it because it would be irresponsible to conceptually remove the entire benefit of end-to-end encryption because it's encrypted. In transit, encryption is not broken. That's what some people came back with. And I'm like, yep, I got that. But if somebody can read the message before you send it and before it's encrypted, or after somebody sends you a message and after it's decrypted.
paul (45:31.428)
Then you defeat the purpose because it's a third party under the Cloud Act. Every tech, every company in the US has to follow US law. If their servers are in the EU, if their data is in the EU and all their customers are in the EU, it doesn't matter. The US government has laws that mandate them to give that information when requested under certain circumstances. And there are certain laws.
That prevent open AI from even telling people when that information has been collected. And that's text messages between lawyers and clients, mergers and acquisitions, therapists. And it's not just what happens from now in the future, it's every conversation anybody has ever had with an Apple customer is potentially on the OpenAI servers and potentially accessible to the US government.
And Edward Snowden has taught the world that governments don't just target specific people with surveillance. They listen and read every message, every call in bulk without a warrant. So if we think these intelligence agencies and governments have all of a sudden got a conscience, I you should look at history.
And I'm familiar with the technology can do because in 1999 I was brought in to conduct an audit on the technical operations of a US company that does lawful intercept. And that's now called SS8. And it's what the six main intelligence agencies, spy agencies, use for lawful intercept. So I'm I'm I know what the tech stack is capable of.
And what we're seeing right now is AI accelerating the agenda of a small handful of people across certain states and tech companies that want to have surveillance and control of populations. So it sounds very dark and sinister. but when you join all the dots and I've just given you I've just scratched the surface. So yes, we are. The Russians have this, I think it was a Russians that.
paul (47:52.443)
came up with a technique called reflexive control. reflexive control I think it is, where people make a decision and think it was their own idea.
We have people lobbying for child safety, not realizing that they're actually begging the government to surveil them.
Francis (48:10.905)
So Paul, th th this has been super insightful and I've been letting you speak because th there's so many different strands that have to be intertwined here to to create that kind of silver tread. But the silver tread is becoming very apparent. I'm gonna get my hammer now and beat the beat Jesus out of my two iPhones. They're sitting on the desk beside me listening to listening to our conversation. But I I have to I have to ask
Wha what's the what's the fix? Is is there a solution? Are we blindfolded walking into the abyss here? Like like I I I have been in security for most of my professional life, and in the last 18 months, I have never seen technology evolve at such a speed with such a lack of consideration for privacy, with with such I suppose enthusiasm to implement things that people don't fully understand.
And it's been embraced with open arms across enterprises all over the place. And my fundamental takeaway is I've I've I've started to step away from chat box, get back out the the books to read and you know, started writing in in in copies of pen and paper because I actually felt I was starting to lose some of my edge, some of my sharpness. And these conversations help me because I'm I'm talking to a human. But but I feel like we've almost become stuck in the scroller.
our our perception of reality. I've never met so many angry people. I've never cross come across so many parents who tell me their kids are suicidal or depressed at ages they shouldn't even know what those things mean. And and now you're kind of bringing the treads together of those facets being brought to a level where they can erode the actual privileges we all have in terms of privacy, in terms of confidentiality and conversations that we've had. And that that erosion can go back
in history now and you know, be pulled out and potentially used against us into the future. So is there a fix or do we need to wake up and start having proper conversations at the right levels across government, across society?
paul (50:17.048)
I don't know if we can fight the tide. What we can do is we can try our best and slow things down. you see, the problem is convenience. Convenience and security have always been at odds. Security and privacy have been at odds. So over the years, depending on whatever it is I'm advocating for at the time, I always find myself fighting a certain sect of people who seem to be quite extreme in their views of a particular thing.
So today, like that's why I I say I'm a technical expert who just happens to be an advocate for child safety, privacy and civil liberties and personal safety, right? Across the board, I'm unbiased, independent. I've been involved in all of those areas. But when I write today, for example, about if I wrote something about the parental controls that are available on phones, there will undoubtedly be people that will remind me of how dangerous the tech companies are.
And it's not the parents' responsibility. And I don't care about child safety. Like, like, like I don't have the expertise to know what's actually in their best interest from a child safety perspective digitally. I would say that we can make a difference. I had a call with a government official from a European state about one or two months ago, following him.
Reading some of my posts. He just happened actually I don't want to say his profession, but he happened to have a very good appreciation for the device controls. I gave after the call, his country, his government was looking to implement identity verification for the purpose of child safety. He persuaded them to instead roll out a
Campaign, an awareness campaign to educate parents how they can actually protect their children. So when I get onto a podcast like this, or interviewed by Fortune or Wired or a Broadcaster or wherever, it's a two-way street. I can give them some insights and perspectives they haven't thought about, but I benefit massively because it gives me the opportunity.
paul (52:40.581)
To make more people aware of something that's important and then people take action. So we need to push back. We need to push back, but we need to join the dots. And I'm I'm starting to do that for enough people at the minute. Like I'll give you one example without without making any conclusions. Ireland has a new
Maritime defense strategy for the first time in history. That was announced less than a year after OnDrill acquired an Irish, Irish-based maritime defense company. Andrill is one of Peter Teal's companies alongside. It's a sister company of Palantir. So when you know when you look at
The war on Gaza and Iran. And without getting into politics, because I've had people say to me I'm the only person they've seen write about Israel and not be called anti Semitic. And that's because I'm not, and it's because I just talk about the technology and security and what's happening. Andril is the company that is the defense tractor with the hardware, and Palantir is the one that directs them and tells them where the targets are. So Andreil has acquired an Irish company. So there are staff here.
There are staff over there. Sorry, I'm actually I keep thinking of in Ireland sometimes. So there are people working for Andrew in Ireland working on a maritime defence tractor con con defence strategy. Ireland is also where Peter Thiel decided to have his dialogue meetup, which included an a person who's responsible for deciding which tech companies are able to take from the grid.
And I could talk about that forever when it comes across different states, whether it's I'll give you another example, and this concerns me and hasn't hit any of the media in Ireland. And I don't just write about Ireland, I write about all states. But Ireland is an important one because it has a very strong presence in the EU when it comes to data privacy laws, and it has a presidency in the EU, and security is one of its things. So in Ireland, the regulator
paul (54:57.997)
Awarded an Israel group, an Israeli group, it's based in Ireland, it's an Irish group. They've awarded that Israeli lobby group or a group, sorry, I don't think they're a lobby group, trusted flagger status, which means they're responsible for deciding what content in Ireland on social media is deemed anti Semitic.
And they have a special relationship with all of the tech companies who must prioritize their reports. And they have already said publicly that Ireland was wrong in everything they have said it has said in terms of Israel. Now, without going into politics, that's irrelevant for me, right? For this conversation. For me, I'm a specialist in content classification. The type of content is completely irrelevant.
Whether it's fake news or disinformation as we call it now, child safety, it doesn't matter what the classification type is. It is the technology and the people involved. You should be unbiased and independent to decide what should be done or what should not be done. And there's lots of other dots to join around how that happened and who met beforehand. I could say the same thing for what's happening in the UK, the US, Canada.
Australia with very specific individual entrepreneurs, specific tech companies, specific intelligence agencies, including 8200 intelligence in Israel, or the intelligence in the US, or intelligence in Australia, or GCHQ in the UK. What we're seeing is this when you talk about this accelerated movement and growth in technology, you're absolutely right. I mean
I've been through a number of different the development of a number of different ecosystems and I haven't like what we're seeing now is absolutely incredible. But it's not just that. I I don't think I've ever seen legislation being created so fast. I've never in my life seen politicians move so fast to roll out legislation and policy decisions involving the protection of children. You know it's
paul (57:21.954)
It's it's it's hypocritical. Ireland has to be forced by the EU to enforce fines on Meta. But at the same time, it says that Meta is causing a lot of harms. Ireland is leveraged also, right? As you know, 40% of Ireland's corporation tax comes from just two tech companies. And I believe, and I could be wrong, but I believe that represents about 50% of the country's tax receipts. So
When those small number of tech companies meet up with the Irish politicians, what do you think they talk about? I think IDA Ireland is absolutely a phenomenal organization. I was one of the first beneficiaries at AOL in nineteen ninety six when IDA started with their inward investment strategy. Thirty years later, those tech companies are leveraged and Ireland is at incredible risk.
So, what will those tech companies do with that leverage? It's not like they have a track record of implementing things that are in the best interests of society. What do you think they might want? Maybe you might conclude that they want AI data centers. And what most what a lot of people don't realize is there's this very significant difference between a regular data center and an AI data center. I'm all for AI data centers.
We're going to need them in the future, but the problem is, and if you read Karen Howe's book or listen to her Steve Bartlett podcast, she's a phenomenal person with great insights. It's the speed by which they're building the AI data centers that are not efficient and they are guzzling, they're being built so inefficiently, they're guzzling energy, they're guzzling clean water. So
It's all connected, Francis. And AI is what's accelerating what I'm seeing. And what I'm seeing is this incredible effort, this incredible coordinated effort to have surveillance and control. And here's one reason. In the UK, 30 people on average are being arrested every day for something they've said online. For something they've said.
paul (59:45.423)
It's usually or a lot of the time about Israel, or it might be just about Gaza. It doesn't matter if it's about Gaza, Israel, Zimbabwe, the US. It doesn't matter. You're saying something that isn't illegal and they're being arrested. So, and then also some politicians get bullied, which is disgusting and horrible, and that should come to an end. Their response force everyone to have their identity verified.
So nobody can be anonymous. Online anonymity is extraordinarily important for freedom of speech and personal safety. It's something that we all deserve as a human right. Nobody gets to ask. When somebody asked me on X, on Twitter, what do you have to hide in private messages? My response was, It's none of your business. And that's the point. So
Francis (01:00:38.33)
Paul, I I'll I'll I love that. We're gonna I'm gonna put that in the in the tagline, you know?
paul (01:00:41.156)
Yeah. It's it's none of your business. And that is the point, right? I don't have to explain to you why I want privacy. It's not for you to ask. And so when everyone has to write have to have their identity verified, that has very dangerous repercussions. Watch this. I had my account on LinkedIn restricted three times for no good reason.
The first time for an hour, second time for eight hours, third time for 48 hours, and then I got on a call with the escalation team and it hasn't been restricted again. But one could argue that the algorithms aren't amplifying my posts like they used to be. Now all most of my posts on LinkedIn are in the top half of 1% for readership and engagement because I write about the stuff that I just articulated and I just do it in a very short, succinct way.
I name names, I say what's happening, I just draw in what's happening, what's been said, and I add a few other things to the recipe, and then you draw your own conclusions. And so people engage very heavily. Now here's the problem. If I got banned for life, I could spin up another account with an email address. People who get banned on Instagram, they could lose their business, but they can spin up another account with a new email address. What happens?
If you have to verify your real government identity, and that's the direction we're going in. Once you're banned, you're banned for life. So you're banned for life from a small handful of companies, which means you've been exiled from from society, from digital society. So what's worse is that when people
Get to realize that where there is a country that does have identity verification. And they do decide it's not just about child safety, it's about illegal content and whatever that might be. Because every government has done scope creep in this area for the last 15 years. What people will do is they will suppress themselves. They will
paul (01:03:05.905)
Be afraid they will censor themselves in what they say. They'll be afraid to say anything negative about a politician, about a government leader, about a tech leader, about a specific state, whatever state that might be. And I'm not trying to be smart by saying that. It could be there's a number of states that wouldn't want anybody talking negatively about them. That is the worst situation you could possibly have where people are afraid.
To say what they have to say online. It it that's what I'm scared of. And you know, when we look at what companies like Palantir are doing, they're rolling out predictive policing. So now we're looking at states, governments, who decide what can potentially happen to somebody based on them making a guess without the person doing anything. In a relationship, if somebody says,
you always do this thing, or you're going to do this thing. The right response is at least give me a chance to get it wrong and apologize afterwards, right? Don't like it's okay to make it an assumption, but try not to act on that assumption. And we're looking in the UK and are rolling out surveillance cameras across the streets of London and inside stores like Sainsbury's, where normal customers who haven't done anything wrong have been barred until it's proven that they actually didn't do anything wrong because of all the false positives and
When AI bans you from Facebook, Instagram, TikTok, LinkedIn, or Twitter, there's no human at the other end. It's not like AOL in the 90s. Now you could say, but there was a lot fewer people online back then, but also there was a lot fewer employees and a lot less technology. So everything is relative. It's not that hard. You know, when my account, let's say, because I had this call with them and I said,
If I have this number of followers and I have this kind of engagement and you've seen the graphs and the charts and I've shown you the data, surely there is a way to give that to the team to learn that there must be false positives. So when you fix it, it fixes it for many. But these companies are so arrogant that they're right, their AI is perfect. You did something wrong, you must have done something, even when I
paul (01:05:31.965)
Prove that when I said that there are LinkedIn threat intelligence people who say they follow my posts, it's the best thing in their feed. Surely that should tell you when I say that I'm not using AI to automate anything, that the chances are I'm telling you the truth. So AI can be used as a scapegoat. So the introduction of this AI slop flagging option.
On websites like LinkedIn that was brought in recently. That could also be a way for these companies to excuse humans pulling levers on certain accounts to suppress things that you might be saying. Because when I write on LinkedIn about Microsoft being found out to have literally helped the Israeli military build
A system to surveil the Palestinian population on their phones, where some of the data was hosted in Ireland. And I believe there's a there's potentially a lawsuit against Microsoft in Ireland for that because the data was in the EU. When I write about that on LinkedIn, I'm just writing about stuff that's already been said and said somewhere, and then I bring it together. So it wouldn't be hard.
To guess that LinkedIn might not want me saying those things. So that's why my readers asked me to set up a Substack. But why? Why can't we just live our lives, try to do something positive for the world? But why do these CEOs and you know many of us are neurodiverse? 35% of us are dyslexic, and many of us have ADHD, which might explain.
The traits that are required to try and get through the tough times and be able to solve problems that might not be obvious, spot patterns that people don't see. But there are also people who have different types of conditioning. And also some of them are just sociopaths. They just don't care. It's a game. And in fact, some of them don't just want to win, they want everyone else to lose because that's part of the winning.
paul (01:07:57.321)
So we need to push back. We need to be more mindful. We need to stop believing everything that the AI companies and governments tell us. When OpenAI, Meta, Claude all tell us that their AI is too dangerous. And then they tell us that they actually didn't configure their environment properly and their partner screwed up.
Every major one in the US uses a company based, and I keep bringing up Israel, but it's based in Israel, the one company responsible for all of this for finding their vulnerabilities, their ex military and unit 8200 intelligence officers. And so some people might not want that. Some people around the world might not want the US company they're using for the most vulnerable things, having all of that tested by.
People that they might not trust. They might see as a foreign adversary, depending on context, depending on timing. And so when they say their AI is too dangerous, when in fact, actually, they just didn't build a proper sandbox, you can't break out to the internet. It's not magic. You gave it access. You didn't not give it access. And then all of a sudden, a few weeks later.
They're promoting the fact that they're selling AI cybersecurity services to protect you from the dangerous AI that they created in the first place. Right? They're starting the fire. They're telling you the fires are dangerous. Then they're setting up the fire service to put out the fires. And they'll probably set up a fire related insurance company should something go wrong. So they're they're like, let's not listen to that.
They don't have our best interests at heart. When bookstores are compla are saying that they're seeing these ridiculous sales of books that they've never seen before. And then on a separate paragraph it's Anthropic actually has technology where they've bought all of these books, scan all the books and then destroy them because if they held on to the copy, it would be illegal.
paul (01:10:21.832)
But if they destroy it, it means it's not illegal. What kind of person or people do that? How can you acquire, you know, limited edition books, consume all that knowledge because I want to own all the knowledge and then I'll license that knowledge back to everybody? Like like okay, I use AI, I use multiple AIs, I use it, I believe, responsibly.
But I use it. It's phenomenal. It is fantastic. But that doesn't mean I trust any of the companies or people running the companies that own the AI that I use. It doesn't mean I trust any of the lawyers that come out with this legislation in the US after Meta and TikTok literally lobbied them for this legislation. They helped write the bill. It it's ridiculous. If an executive is found to have caused
The death of other people intentionally through the amplification of harmful content and then try to cover it up and lie about it.
And they're still hanging around with the politicians in the EU. They're still having these selfies talking about how to scale and remove the red tape. That will tell you the true agenda. So that's very bleak. That's very sinister. it's but it is what's happening.
Francis (01:11:50.231)
Very, very bleak. Very sinister, Paul, but thoroughly inj enjoyable to listen to from from from from from my perspective. And and I think there's a huge amount of information there that people will have to have to process. I know we're I know we're up on time, but it's been an absolute pleasure having you on. Really insightful, really, really thought-provoking conversation, and I'm sure the listeners will have lots to say once once this goes out.
paul (01:12:16.028)
Yeah, half of them would be like, You could have said that with fewer words. That's unfortunately one of the downsides. well the pa one of the positives about dyslexia is pattern matching and and whatnot. But also one of the downsides of ADHD is sidetracking, going down rabbit holes and and just talking forever. So I apologize to those that I've talked their ears off, but if anybody's gotten this far, I really appreciate their time and attention because we we know that.
Life is hard and you know, time is precious. We can't get it back. So appreciate anybody who who who stuck it out.
Francis (01:12:55.066)
Exactly. No, it's great stuff. Paul, thanks a million and I really appreciate you coming on the show.
paul (01:13:00.137)
Pleasure.
.png)
