AI Wants Your Life: Tech Boss Meredith Whittaker Says No

19 Jun 2026 · 45 min · 22 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Meredith Whittaker argues that privacy and encryption are essential for human rights, journalism, and even military strategy, and warns that AI “agents” and government device-scanning proposals could undermine Signal’s end-to-end encryption. She also critiques surveillance-based business models and urges systemic governance of powerful technologies.

Guest backgrounds

Meredith Whittaker is a former Google executive/“tech boss” and privacy advocate; she helped build Signal, a nonprofit encrypted messaging app with open-source Signal Protocol (developed in 2013). She studied literature and rhetoric.

Key claims

(1) Signal protects message content and metadata (“who/when/with whom”), collecting “as close to no data as possible.” (2) Meta/WhatsApp’s data collection plus AI enables surveillance advertising and potential subpoenas. (3) Encryption can’t have “backdoors for good guys.” (4) UK/EU nudity-scanning or mass device scanning is dangerous mass surveillance and likely “phones home.” (5) AI agents with broad app access create a “backdoor” into Signal.

Notable examples

“Signalgate” (journalist added to a classified Yemen chat); Snowden’s 2013 use of Signal; a cited case where Meta turned over Messenger DMs leading to a US jail sentence; UK under-18 nude-image scanning proposal; discussion of Microsoft Copilot “AI Christmas” requiring access to Signal and other apps.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Meredith Whittaker's Unique Perspective

0:55 to 1:18

Learn about Meredith Whittaker's background and her approach to privacy.

“Being a small business owner isn't just a career, it's a calling.”

Meredith Whittaker's Unique Perspective

2:38 to 3:41

Learn about Meredith Whittaker's background and her approach to privacy.

“Whether it's social media bans, AI and military strikes, the destruction of jobs, debates about tech in our lives are all around us every day.”

Signalgate and Its Implications

3:41 to 4:30

Understanding the incident known as Signalgate and its relevance.

“And also, it has a deep, and she says, unshakable promise to barely even collect our data, let alone sell it.”

UK Government's Plans on Privacy

4:30 to 5:00

Discussion on the UK's controversial plans about underage privacy.

“As you'll hear, she's got views on that, both the principle and whether the idea is even workable.”

Meredith's Journey in Tech

5:00 to 6:36

Explore Meredith's shift in perspective about privacy during her time at Google.

“Obviously, I think some people listening to this might be pretty new to Signal, but I think anything that you can say to let us into your world and to tell us about you as well as what you do, I think would be great.”

Ethics in Technology

6:36 to 9:26

Meredith discusses the ethical responsibilities of tech companies.

“I did not get a Commodore 64 when I was a kid and everything took off from there.”

Signal's Core Promise

9:26 to 11:47

Understand how Signal ensures privacy through end-to-end encryption.

“How do we determine where it's used, where it's not used, how it's developed, and what the goals of that development are?”

Comparison with WhatsApp

11:47 to 14:00

Delve into the differences between Signal and WhatsApp regarding data privacy.

“And they also apply it to many of their There are messages that are sent.”

Surveillance Advertising and Privacy Concerns

14:00 to 17:35

Explore the concept of surveillance advertising and its implications on privacy.

“Meta is a massive data collecting entity.”

Surveillance Advertising and Privacy Concerns

17:36 to 18:15

Explore the concept of surveillance advertising and its implications on privacy.

“Lately, it feels like there are two types of investing platforms.”
Show all 22 chapters

Surveillance Advertising and Privacy Concerns

18:22 to 18:38

Explore the concept of surveillance advertising and its implications on privacy.

“Brokered services by Public Investing, member FINRA SIPC.”

Signal's Role in Privacy and Security

19:46 to 28:00

Discuss the implications of Signal's use in sensitive communications.

“2013, when Edward Snowden used it and trusted it to talk to journalists about US government surveillance.”

Understanding Technology's Limits

28:00 to 29:20

Explore the inherent limits of technology and its implications for privacy.

“Well, part of being a big brain in tech, I think, is really understanding the nature and the limits of what technology can do.”

The Necessity of Privacy in Communication

29:20 to 31:30

Discuss the essential role of privacy in communications and the risks of backdoors.

“The EU also had temporary legislation which just ran out, but it was essentially asking tech companies to voluntarily scan their platforms.”

Funding and Sustainability of Signal

31:30 to 32:50

Learn about the funding model of Signal and the challenges it faces in the tech industry.

“I think we need to just recognize that, yes, you have people who have a very vested interest in infiltrating those networks, in accessing that communication, in accessing that data, and will spend resources doing it.”

Concerns About AI and Data Privacy

32:50 to 34:20

Examine the implications of using AI tools and their potential risks to privacy.

“One of the real issues we're working with is that the economic engine of the tech industry as it's shaped now is surveillance.”

The Future of Autonomous Agents

34:20 to 36:30

Discuss the impact of autonomous agents on personal privacy and decision-making.

“is in the hands of the entity that is running that service.”

Navigating Privacy Threats from Big Tech

36:30 to 41:30

Understand the threats to privacy posed by centralization in tech and advice for users.

“And I say, hey, Copilot, I don't want to live my life.”

Skepticism Towards AI and Data Use

41:30 to 42:00

Encourage a skeptical approach to AI and the use of personal data in technology.

“these devices, back into these centralized companies.”

Skepticism Towards AI Technology

42:00 to 45:50

Learn about the importance of skepticism when engaging with AI technologies.

“I do come from maybe an old school where I think you should verify then trust.”

Personal Choices and Tech

45:50 to 47:26

Explore how personal choices shape tech engagement and societal impact.

“Not simply if I'm given the choice yes or no, which do I choose?”

Tech Industry Insights

47:26 to 48:48

Gain insights into the tech industry dynamics and its influence on society.

“You can see a photograph of her in front of the famous black door and get all the wider context if you read the written version of this at bloomberg.com forward slash Michelle.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. Never bet against American grit or American energy. Through innovation, Venture Global is not only building some of the largest energy facilities in the world right here in the United States, but delivering American energy at a fraction of the cost and a fraction of the time.

0:48So while others are busy talking, we're busy building. That's Venture Global. That's unstoppable energy.

1:00Being a small business owner isn't just a career, it's a calling. Chase for Business knows how much heart and effort go into building something of your own. Manage all your business finances, from banking to payments to credit cards, all in one place with Chase's digital tools. Plus, access online resources designed to help your business thrive. Learn more at chase.com slash business. Chase for Business. Make more of what's yours. The Chase mobile app is available for select mobile devices. Message and data rates may apply. JPMorgan Chase Bank N.A. Member FDIC. Copyright 2026. JPMorgan Chase and Company.

1:38Bloomberg Audio Studios. Podcasts. Radio. News. We can't live a full, dignified life with the ability to question authority, to test ideas, to blow off steam if we don't have private communication. You can't do military strategy. You can't do human rights work. You can't do journalism. Full stop. Meredith Whitaker, tech boss and defender of privacy. I got the impression that by the end of your time at Google, you were kind of a thorn in management side. It wasn't Meredith versus Google. There is a groundswell of sentiment within tech now that is concerned with the fundamental question, how do we make sure that we are collectively governing this powerful technology so that those who would do harm with it aren't those who are in control?

2:33From Bloomberg Weekend, this is The Michelle Hussain Show. I'm Michelle Hussain.

2:42Whether it's social media bans, AI and military strikes, the destruction of jobs, debates about tech in our lives are all around us every day. I'm old enough to remember the 90s and the emergence of what today we call big tech. Back then, the companies were perceived as benign. They were full of important and exciting features that would connect us to people around the world. And they were pretty much free to use. Today, the story is much more complex. But even so, how many of us really think hard enough about our choices in tech, the products we use, and what the companies get out of us? Well, this conversation is with Meredith Whittaker, a tech boss who is different from her peers, partly because her product, the messaging app Signal, is non-profit.

3:41And also, it has a deep, and she says, unshakable promise to barely even collect our data, let alone sell it. If you're not a Signal user, but you know you've heard about it somewhere, It's probably because of the moment last year when the then US National Security Advisor added a journalist to what should have been a classified group chat about military action in Yemen. It became known, of course, as Signalgate. So that's some of the background. But as Meredith Whitaker came into the studio, there was a live issue in the UK. the government's plans to make it impossible for under-18s to take or share nude images on their phones.

4:30As you'll hear, she's got views on that, both the principle and whether the idea is even workable. But I had questions too on whether everyone has a right to be private. What about crime, abuse, terrorism? Ultimately, I hope you find this as compelling as I did. Meredith Whitaker is able to see the ironies in her world, to spot the moments of humour even, and at the same time, make us think. Obviously, I think some people listening to this might be pretty new to Signal, but I think anything that you can say to let us into your world and to tell us about you as well as what you do, I think would be great.

5:17I'll do my best. I think Signal's a bit more interesting than I am, but we'll see. Meredith, I guess I would just love to start by asking about your own journey towards what you do now. Because you worked in big tech at Google for more than a decade. Was there a moment in that period where you really started to think differently about privacy and data? A moment you can identify that's kind of critical in the development of your story? I have considered that question. Is there a point where I tripped over something or a revelation happened? You know, Athena was birthed from the forehead of Zeus and suddenly there's insight.

5:58I don't think it was a moment. I come out of a non-traditional background for somebody working in the tech industry. I studied literature. I studied rhetoric. I was always an arts kid. And that's still where I as a human being really live. I love it. And so I was, I think, calibrated to perhaps ask different questions or just pursue different inquiries. But I think it was over time beginning to understand Google's business models, the political economic forces that were animating the company. After coming in as basically a novice, I did not pursue computer science in school. I was not interested in tech.

6:44I did not get a Commodore 64 when I was a kid and everything took off from there. And I loved computers, right? The classic origin story. And so I had to ask really, really fundamental questions. Like, what is a server? Would I be right in thinking of you as a bit of a contrarian or maybe a rebel? I know that there was a key moment in your time at Google where you organized colleagues against a Pentagon contract that you realized was being signed? I wouldn't adopt those terms myself. You know, I am somebody who thinks it's really important to maintain the dignity of my analysis. Like if something is true, if there is empirical evidence supporting a conclusion, if we can look at the map and clearly see that there's a problem on the road ahead, then I am not somebody who's going to go along to get along with a convenient delusion.

7:44Those things, once I see them, I'm not going to veer away or pretend that they don't exist, even if I have to take steps that might be scary or against the grain to do that. I got the impression that by the end of your time at Google, you were kind of a thorn in management side. You were organizing walkouts on other issues, ethics issues, and they saw you as a problem? Well, people who would like to put the goals of infinite growth, infinite revenue, profit above the reality of the harms or the problems or the choppy seas on the road ahead don't like people who point that out. But I do want to sort of frame that it wasn't It wasn't Meredith versus Google.

8:33This was a groundswell from thousands of people within the company. Yeah, there were a lot of the walkouts at the time where people came out in big numbers to protest at different things. We see that echoed today. DeepMind's employees just unionized, concerned about the application of that technology for harmful purposes. You have open AI employees beginning to organize or that's being reported in preparation for the IPO. There is a sort of groundswell of sentiment within tech now and outside of tech that is concerned with the fundamental question that I don't think can be avoided by anyone. How do we make sure that we are collectively governing this powerful technology so that those who would do harm with it aren't those who are in control?

9:25How do we scrutinize it? How do we put guardrails on it? How do we determine where it's used, where it's not used, how it's developed, and what the goals of that development are? Those are fundamental questions that every world leader, every government, every community considering a data center needs to be looking at and where the leverage for governance, where that fulcrum exists, is an open question given the bipolar tech world we live in and the volatile geopolitical environment we're in. And the fact that the tech world seems to be becoming ever more powerful, not least with the amount of interest in the IPOs.

10:00Absolutely. I want to take a moment to really understand what you do at Signal. The core promise is end-to-end encryption, which means only the sender or receiver can see the actual content of the messages. So how is that different from WhatsApp, which also has the same promise? Well, the core promise of Signal is that in the digital world, where almost every micro action is surveilled, turned into data, processed in models, sold to advertisers, sold to data brokers, et cetera, et cetera, we preserve the fundamental human right to private communication, full stop, which has existed for hundreds of thousands of years.

10:42That was the norm. If you and I are talking, that conversation is gone once we are done talking and part ways. And we have moved into a world where that ground has shifted dramatically out from under us. And that is no longer true. And we think that, you know, it is considered a fundamental human right for a very good reason. We can't live a full dignified life with the ability to question authority, to have different relationships with people, to test ideas, to blow off steam. If we don't have private communication, you can't do military strategy. You can't do human rights work. You can't do journalism, full stop.

11:19So Signal provides that in as robust a way as we can. We developed the encryption protocol, the Signal protocol that moved the field forward in 2013. We introduced that to the world. And that is the core of our offering that protects the message content on Signal. What you send only goes between sender and receiver. Only the people you're talking to know it. WhatsApp licenses that protocol from us. And they also apply it to many of their There are messages that are sent. WhatsApp for business is excluded from this in many cases, but they protect message content. So what you type, what you say is end-to-end encrypted.

12:03They can't see it, only the sender and receiver. However, you have to look at privacy beyond simply the message content. And this is where there are very fundamental differences between WhatsApp and Signal. Signal also applies encryption to protect what is called metadata, But that is extremely intimate. It's extremely revealing data. It's your profile photo, your name, who you talk to, when you talk to them, who's in your groups, how often you chat, when did you text your oncologist, when did you text your therapist, when did you text your lawyer, when did you text someone or your best friend, all of that information.

12:41So WhatsApp does collect that. We don't. We go out of our way to collect as close to no data as possible. And we make that open source. Is it really that revealing, the other data, the message data? If someone can see that I wake up at six in the morning and I send a lot of messages at that time, and they might even know who I'm messaging, my mother, my colleagues, people in other time zones. Is that really so revealing of me? Oh, it absolutely can be. Are you messaging your FBI contacts because you want to be a source? Are you messaging a whistleblower attorney because you found corruption in your workplace and want to report that?

13:29Are you messaging a divorce lawyer? Are you messaging an oncologist? Yeah, but what I mean is, again, it's like WhatsApp, for example, could see, oh, that other person is an oncologist or that other person is in law enforcement or is a source. But again, they can't see beyond that what the actual content is that I'm saying. Well, no, assuming that's not on WhatsApp for business. However, I think we have to expand beyond simply the application and look at WhatsApp being owned by Meta, the parent company. Meta is a massive data collecting entity. Surveillance advertising is the heart of their business model.

14:11So, okay, you messaged your oncologist. All right, let's join that data with all of the other data we have in hand. We buy credit data from data brokers. We have Facebook data. We have Instagram data. And we have powerful AI models that can find patterns and search across this data. So I might see ads for cancer drugs on Instagram? Is that what you're suggesting? Because surveillance advertising is a very inflammatory way to put it. Well, I mean, it's a technical descriptor, right? It's targeted advertising based on the collection of data that reveals or purportedly reveals certain qualities of you and what you might like.

14:53Let's take another example. You're messaging a whistleblower attorney and it's very high risk. And the corruption that you found is extreme. So they know, okay, this is whistleblower aid, an organization that helps with that. Okay, it started messaging now. Okay, this is the contact. okay, well, what could they do with that? What could they reveal? Could that be subpoenaed? Could that be revealed to your employer? The fundamental answer is that yes, this data is extremely revealing. It reveals patterns about you. We could hypothesize, but we know already that there's a woman who went to jail in the US because Meta turned over DMs on Facebook Messenger between her and her daughter that showed her helping her daughter access and deal with reproductive care in the state of Nebraska.

15:50Do you find it frustrating then when you see people not care about their privacy? Some people wouldn't even be aware, but people who are aware, but just think this is the price I pay for these platforms being free at the point of use. or maybe younger people, especially, who just think that there's nothing really in their life that they would mind being in the public domain. Yeah, I find it frustrating when the argument that no one cares about privacy is leveraged by these firms to justify what they're doing. I don't actually find it frustrating when people who really, in my view, have little choice say, you know, hey, I can't do anything about this.

16:36I'm going to delete that from my long list of worries because I don't feel that I am empowered to act on this. I actually don't believe that people don't care about privacy, but I believe that the stakes are very obscure. And we're all living in a world where we have dishes to do. We have bills to pay. We have things that are right on our doorstep to deal with. And thinking about data flows and privacy legislation is probably pretty far down on our list of concerns.

17:10The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. Support for the show comes from Public. Lately, it feels like there are two types of investing platforms. Some are traditional brokerages that haven't changed much in decades, and others feel less like investing and more like a game.

17:52public is positioned differently it's an investing platform for people who are serious about building their wealth on public you can build a portfolio of stocks options bonds crypto without all the bugs or the confetti retirement accounts yep high yield cash yes again they even have direct indexing public has modern design powerful tools and customer support that actually helps go to public.com slash market and earn an uncapped one percent bonus when you transfer your portfolio That's public.com slash market. Add paid for by Public Holdings. Brokered services by Public Investing, member FINRA SIPC.

18:28Advisory services by Public Advisors, SEC Registered Advisor. Crypto services by ZeroHash. All investing involves risk of loss. See complete disclosures at public.com slash disclosures. Whether you're planning a big tech event, launching a new campaign, or just stocking up on team gear, finding the right promotional products makes all the difference. 4imprint offers thousands of options. from on-trend apparel and premium drinkware to tech, totes, and giveaways, so you can find the right fit for any audience, purpose, or budget. You can customize it all, your logo, your message, your look, and many items come with no setup charge to help you save.

19:04And if you're really watching the bottom line, you'll find standout choices at every price point so you can make a real impact while staying on budget. Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee, so you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint. 4certain.

19:45You had two big moments in the growth of Signal, both of which I remember. 2013, when Edward Snowden used it and trusted it to talk to journalists about US government surveillance. And the second was SignalGate last year when a journalist was added to a supposedly highly classified Trump administration discussion on military action. When the second happened, what did you think? This is like a great advertisement for Signal because it's secure enough for the U.S. government? I thought, whoa, what's going on this morning? This is an interesting story.

20:26The article hit my feed the same time it hit anyone else's feed. And I think I had to read it twice to really let it sink in because it read almost like a telenovela. What I didn't anticipate after reading it initially was that Signal would become a main character and that it would become a story about Signal. And what I was primarily concerned about is that suddenly Signal was being reported on, spoken about by millions of people who weren't familiar with the intricacies of encrypted messaging or the technology itself. And I wanted to prevent misinformation that could hinder signal or misinform people who, you know, often need to choose signal in life or death consequences.

21:16Because this was essentially about human frailty, someone who shouldn't have been added to that chat being added. And your worry was that it would make it look like signal itself was insecure. Yeah, it would be framed as a problem with signal or a problem with privacy or, you know, that was a very volatile, very heightened news cycle. So how that would turn out was not clear. But yes, I mean, this was ultimately men's thumbs doing things. That is not something an application developer can take responsibility. Was it good for numbers? It was, yeah. It turned out well for us. We saw growth across the U.S., across the world, and we've seen pretty steady growth since then.

21:53Can I ask you, Meredith, how you reconcile yourself with where your product is being used by people to do harm to others? And particularly the example of child sexual abuse. Am I right in thinking that if pedophiles, abusers are using Signal to circulate images of child sexual abuse, that can happen without them fearing that you will hand over data to law enforcement? We go out of our way to not be a social media company. So we could pump our numbers fairly easily if we added things like broadcast channels or features that allowed you to find people with similar interests and you meet strangers on Signal.

22:38But Signal has made a very conscientious choice to stay in our narrow lane as an interpersonal communications app, which means you have to encounter the person you're talking to in some capacity that is not a viral find a friend feed. Yeah, that's outside the app essentially. But once you do find that person, you can be sure that anything you share with them is free from outside surveillance. Yeah, that is Signal's Promise. We are open source. People audit our code regularly. We have the open source community looking at our code. We cannot collect that data. We cannot turn that data over. And look, we limit the features that allow virality.

23:26We limit the features that allow mass distribution. And we do that very conscientiously. I also need to make a fundamental point that is, you know, it is simply true that encryption either works for everyone or it works for no one. There is no way to have a golden key or a special backdoor or access just for the good guys. And do you ever struggle with that? Knowing that there will be people who are using it for the most abusive possible images of children who are on signal right now? Like, Are there moments where you find that hard to reconcile or come to terms with yourself? The fact of child sexual abuse is horrifying to me.

24:11Absolutely. And I think we need to take that so seriously as a society that needs to get much, much better with looking gnarly problems in the face and actually tackling them. But I also struggle with the persistent focus on privacy as the purported barrier to solving these issues when we have most of the Epstein list walking around without consequences, when we have so few mechanisms to actually help children when they report a powerful figure in their community, when we are defunding the actual law enforcement officers who take this data, act on this data, move to prosecute the uncles, the fathers, the pastors who are perpetrating this abuse.

25:04Can I give you the example in the UK where many people would think you have to use every tool in the block, right? So you have to use the levers you're pointing to, and perhaps you have to use the levers in tech as well. So the UK government is moving towards making it compulsory for tech companies to ensure that children can't take or share nude images through their platforms. Is that workable, do you think? And if the UK compelled that, what would it mean for your presence here? Well, I'll start by saying there's never been more data accessible to law enforcement than there is now. That were law enforcement well-funded, were social services well-funded, they could act on and often don't, often delay or often simply don't have the capacity.

25:54So when we talk about every tool in the toolbox, we need to recognize that there are some tools that are not being picked up. And there are some tools that seem an obsessive focus of politicians who like to abstract these very difficult problems into the technical domain and pretend there is one technical fix they can apply to solve a problem that needs a social solution. What do you think of it as a policy proposal? of mass scanning of devices that would be age-gated. We put out a statement. Keir Starmer announced what seems to be a fairly hasty proposal to implement nudity scanning as a mandatory feature on all devices that would be sold or used in the UK and that you could only opt out of this nudity scanning if you provided some sort of verification that you were above a certain age threshold.

26:49Correct. There are very few technical details on that system, but our position and our position on the client-side scanning debate that circled the online safety bill as well, now ACT, is that this, you know, this constitutes very dangerous mass surveillance, that having a scanning system on your device that is going to, you know, purportedly detect nudity can easily be weaponized by governments for political speech, for other content, and that we don't understand the technical mechanism by which this would be implemented, because my assumption is there is no specification for this, but that we would be looking at a system that would likely phone home.

27:37So that's sending that data to governments, to tech companies, and ultimately comprise a mass surveillance system of permitted speech that would be implemented across everyone's device and pose real threats to Signal's ability to provide the robust privacy guarantees that militaries, governments, Ukraine, human rights workers, journalists, people around the world rely on often for life or death. With all of the big brains in tech, you really don't think it's possible to refine the product in a way that protects children and does the age verification and the certain kinds of images without it having these massive repercussions and implications that you're describing?

28:22Well, part of being a big brain in tech, I think, is really understanding the nature and the limits of what technology can do. And this isn't a question of go think harder nerds, right? There are real limits to what tech can and can't do. And I think tech's coming up with new things all the time. Their marketing departments are talking about it all the time. But if you look at the implementation layer, often what you're looking at is sticky tape and twine and a handful of legacy libraries that may or may not be suitable for purpose. The limits of your software, the limits of the memory that's running on your hardware, these are fundamental material limitations.

Read the full transcript

29:06And there is no technology that has been shown to do this in a way that is robust and safe. Yeah. I also just wondered how sustainable your position is. I hear what you're saying, but it's not only the UK government that's really concerned on this front. The EU also had temporary legislation which just ran out, but it was essentially asking tech companies to voluntarily scan their platforms. And we've taken the same position on that. this does seem to be the way that governments are going. And if it is, can you imagine Signal not being available in the UK or the EU? Like you threatened to walk over the online safety bill as it was then in the UK.

29:45Look, we don't really do PR, right? We have a position and our position is we would rather exit a market than undermine the technical guarantees that people trust for their privacy. That doesn't change based on jurisdiction. It doesn't change based on year. And that's not something we take lightly. We're not just like flexing or being recalcitrant. That is recognizing that encryption requires a network effect, that a backdoor in any market compromises a door into Signal that would be weaponized by nation state adversaries. Signal is the only mass communications platform that provides the level of privacy that we provide.

30:33And we know that we have adversaries that are consistently searching for vulnerabilities in our code. Who do you mean? Who do we mean? Any government with a capable intelligence agency is spending some of that effort looking at Signal. Including your own government in the US. I don't have clearance to know that for sure. But I'm just wondering if you mean Russia, China, or do you mean governments in the West? We know it has been represented to us that, you know, Russia is, we've seen bounties for the delivery of vulnerabilities, meaning that, you know, they'll pay you if you can find a vulnerability they could exploit that have been offered by the Russian government.

31:15We have had people who would know this or represent that to us. But if you look at the people who rely on this, heads of state, government officials, military strategists, investigative journalists, people reporting corruption, I think we need to just recognize that, yes, you have people who have a very vested interest in infiltrating those networks, in accessing that communication, in accessing that data, and will spend resources doing it. That's the objective function of an intelligence agency. And Signal is the only one out there doing this. So it would be very serious. It would be an absolute breach of our duty of care.

32:02Were we to implement a backdoor in Signal, were we to undermine that promise? And, you know, we're a nonprofit. We are not a tech company that is, you know, juicing numbers to IPO and exit. We take this very seriously. That's one of the intriguing things about Signal. And I wondered how hard it makes it because you have some big donors, but you also have users who donate to you. So today, what are your annual running costs? What are the costs you have to cover? It's about$50 million a year, 5-0. And do you fall short in donations? We're sustaining, yeah. It's about 70 % large donors, 30 % people who use the app and donate$5 a month,$20 a month.

32:46We love them. Thank you. Does it feel like a struggle?

32:52It's something I think about a lot. One of the real issues we're working with is that the economic engine of the tech industry as it's shaped now is surveillance. It's the collection and monetization of data. That's not alarmist. That is technically true. And so existing within a framework where all of the pressures for revenue, for growth, for profit, for IPO are pointing in the direction of collect and monetize data means that we have to choose a different model. One where we don't have a VC saying, hey, you're not going to meet your quarterly goals, you're not going to get your Series B if you don't figure out how to monetize, where we don't have a board and shareholders saying, figure out how to carve out some data collection so you can build an AI model, so you can continue to meet and exceed market consensus.

33:46Well, that made me wonder about the LLMs and ChatGPT and Clawed. What would you like us to be aware of in privacy terms when we use those, when we ask questions, when we put prompts into the chats? Well, I mean, flatly, like these are not your friends. These are not conscious beings. These are not sentient interlocutors. They're very useful. Sure. Many things are very useful. And you can delete chats and you can say you don't want your chats to be shared. You can say all sorts of things, but ultimately the power to determine whether that happens or not is in the hands of the entity that is running that service.

34:27And those can be stored. Those can be logged. Those can be mined for more data. Those can be tweaked slightly to calibrate the responses to your purported preference on behalf of, say, advertisers. You know, maybe there's an airline that has paid to advertise. And instead of showing ads, perhaps, and I'm working with a hypothetical here. We don't have evidence of this, but this is very much in the realm of the possible. You say, hey, ChatGPT, find me a very cheap flight. And instead of finding you the cheapest flight, it finds you one with their preferred advertising partner that is not cheaper.

35:05Do you use them? I mean, I ask them to format a document here and there, but no, I have no relationship with them. I don't ask them questions. I'm very serious about my thinking and my writing, and I don't want the process of working through an idea, the struggle of not knowing and then figuring it out to be foreclosed or eclipsed by the response of a system that's kind of averaging what's already out there. Yeah, I guess that's a different argument. That's also about safeguarding your headspace or your thinking time beyond privacy. The next generation or the next layer of all these models is the more autonomous personal assistant type thing.

35:53You know, Microsoft Co-Pilot, Mustafa Suleiman was on this program and he said to me, you know, by next Christmas, you can be buying your Christmas presents through Microsoft Co-Pilot working out what you need to buy for whom. What do you think of that kind of autonomous agent being in people's lives? What would it mean in privacy terms? Yeah, I mean, that's a great question. And let's take Mustafa's little vision of, you know, AI Christmas. I can't wait. Yeah, everyone's getting a gift certificate to Microsoft Co-Pilot.

36:30But what would that mean, right? So we have a Windows laptop. It's running Copilot. And I say, hey, Copilot, I don't want to live my life. I would like you to live it for me. And I don't really want to think about the warm feelings I have for those that I'm buying gifts for or think about what they might actually love or engage in that generous interpersonal communal feeling that maybe should be at the heart of Christmas. Who knows? But I don't have the time for that. I'm going to play pickleball or whatever the thing I am doing. And I would like you to do that for me. And I would like you to figure out what gifts to buy.

37:06I would like you to consult with my siblings group chat and see, you know, what our family might like. I would like you to buy those gifts and I would like you to figure out how to get them to the right place. Well, okay, that would need access to my credit card. That would need access to my browser to pretend it is me to drive through that browsing history to find those things. If it's accessing my siblings' group chat, that would need access to my signal. That would need the ability to message my siblings on my behalf. And then it would need my home address, say, to ship those gifts. It would need maybe my calendar to make sure they get there on time, etc., etc.

37:50So what you've just described is a system with very, very pervasive access across multiple applications and services. And importantly, in the context of Signal, it would constitute a kind of a backdoor into Signal. It would be able to access Signal, pull data out of Signal. Your objection isn't just privacy and access to our data. It's actually that you think once something like that ends up on a phone because we've given it permission to, it could actually be a threat to Signal's ethos and promise. We have described it metaphorically as breaking the blood-brain barrier between the operating system and the application layer.

38:33So you can make Signal secure enough to protect against that? Well, what we are seeing in the case of these agents are promises like this, right? The access of data at the application layer, pulling that data into a context window, but sort of a big data slurry that the agent sort of acts on and remembers. And then implementations, architectures for this agent that are very unsafe from a cybersecurity perspective. Because the agent is essentially becoming the person or the agent has been given the license to be me or, well, not you, because I'm sure you would never go down this route. I try not to.

39:12That is my struggle. But you've given it in sort of technical terms, like root access to your life. You've given it the keys to the kingdom because to act on your behalf, it needs to be able to do everything you would do. Do you worry then about this as an existential threat to Signal? It's clearly not that easy to run the foundation. I do. I do worry. Because it seems quite likely that it will come to pass. Just if you look at how much bigger these companies are getting by the day and they're going to need to make more promises about what they can offer. I use the example of Copilot, but there are others clearly.

39:48Any one of the operating system vendors that decides to implement an agent with these capacities is ultimately hollowing out the ability for Signal and others to guarantee privacy or data protection at the application layer. And this is a quiet but profound shift that we are witnessing. Can you fight it? Oh, I mean, I don't believe in inevitability and I am willing to fight on my principles. I don't know if I can win it, but I will do everything I can because I think this is fundamental. I am concerned that we are seeing extremely hasty decisions made by very powerful actors. You know, we have three companies in the world with control over our core operating systems.

40:37That's Google with Android, Microsoft with Windows, and Apple with iOS, macOS. We all are relying on one or another of those. And they can make decisions by fiat that fundamentally harm our collective cybersecurity, that take away any choice we might have for privacy, and that completely hamstring developers at the application layer who've spent so much labor, so much money, so much intention on building different services that protect or do X, Y, and Z. So we are seeing, ironically, almost in the name of AI agents, a move of agency away from the application layer, away from the user, away from the consumer who uses these devices, back into these centralized companies.

41:34Give us some advice then, because I think I could take away, okay, wherever you have the option to tick the box saying to limit the use of your data, do it. But I think you're also telling me that you don't really trust in what you get back for saying, don't use my data, I don't want it shared. Well, you know, look, I think we have evidence that, you know, simply trusting is not a sound conclusion. I do come from maybe an old school where I think you should verify then trust. What is the advice? I think healthy skepticism and a recognition that this technology is not magic. It can be extremely useful.

42:13It can have profound applications, particularly where finding patterns and large amounts of data is useful for X, Y, Z. but it runs on hardware. It runs on software. When we talk about AI, we're not talking about a magical sky god that can suddenly solve all our problems, answer all our questions. We are talking about a statistical model that is then embedded in a hardware software system using software libraries, implementing choices on the user interface level that are made by people that are susceptible to vulnerabilities. Tell me what you do in your own life because Because signal for messaging, what else?

42:51What is common in the rest of our lives that you never use and you think we should? Like when you go for a run, do you put it on Strava? You would not do that? No, I don't. I mean, I'm not a, I'm not like the fulcrum of leverage for me is not individual consumer choice. And I am not trying to be a holier than thou purist, right? You know, I have a tracking app I use for different health metrics. It's been helpful for me. I travel a lot and it's been helpful for me seeing that. That's health data. Yeah, that is health data. That'll be used. Yeah. And I benefit from that. I think the lever we need to be pushing on is much more systemic change.

43:33I am not going to opt out of life. I am not going to go live in a cabin in the woods and like pay someone to print out my emails at the post office in town or whatever we're talking about, you know, so that I can feel private, right? Like if we look even at signals, privacy promises, we're not talking about sort of isolation from our social, economic, political lives, living in a vacuum of like holier than now cybersecurity purity. We're talking about the ability to talk to other people, to share with them. Right. And so, yeah, I have an Instagram. I post mainly work events or whatever on that.

44:09Right. I use social media. I have the NTS app. It's a really good web radio that has great mixes. Right. Like I listen to that. Maybe they collect data. You know, I push reject all, but I don't always trust that that's actually rejecting cookies. I am a person living in 2026 and I want to participate in the world. What I don't believe is that participation should be on the terms of a handful of tech companies located in the U.S. whose choices are made based on a desire to maximize revenue. And yet you too are American tech. The irony abounds. Look at this. A complex individual. I am. And I worked in tech for 20 years.

44:47I'm of the tech industry. So those two things are true. But I think you're saying to us, make better choices or just think more carefully about your choices. Like even if you get that much from people, will you consider it valuable or even a victory? I would say those things and that you be skeptical of these claims. You're not stupid. You know, the choices offered, yes, no, or maybe, maybe aren't the choices you want. As a dream a little bigger. What do you want from these companies? What world do you want to live in? what world would be the thing you want to help build? And then how do you think about demanding from the people who have commandeered, who have created, who now hold so much power over our social and political and economic institutions via the tech infrastructures, the information ecosystem on the tech platforms, the AI decision-making that is inoculating so many functions in our lives.

45:46What do we need from them to help create that world? Not simply if I'm given the choice yes or no, which do I choose? And that's the limit of my ability to enact any change in my life. But do we want tech in our education system and how? Yes, we do. Okay, well then, what does it look like? And how do we work to guarantee that we're actually using it to shape that better world and not simply being shaped by decisions made on a whiteboard in Mountain View by a handful of product managers who really have to ship AI this quarter or they're not going to get promoted. Who you probably know because you once used to work with them.

46:24Yeah, and they're all named Mike.

46:29Even now? Is it still like that? Even now. Even now. I've heard. I've heard. You never miss that world? I mean, it's a very big bucks world apart from other things. Am I sad I'm not like rich rich? Yeah. No. No. Like I'm hood rich. I'm happy being hood rich. I don't need to be rich rich. What does hood rich mean? It means rich in the context of the neighborhood you come from. Of your hood, right. Not rich in the context of complaining about the champagne on your private flight. And your hoods are both New York and Paris. So you get the... I'm like, don't cry for me. Meredith Whitaker, thank you so much.

47:07It was great. Thank you.

47:14And that is where we said goodbye. Unknown to me, later that day, Meredith was visiting Downing Street to talk to Keir Starmer's team about Signal's concern at their plans. You can see a photograph of her in front of the famous black door and get all the wider context if you read the written version of this at bloomberg.com forward slash Michelle. A few other notes. We asked Meta to comment on her characterization of its business model and approach to privacy. A spokesperson for WhatsApp responded by saying this. We also look at privacy beyond encrypting messages. We don't keep logs of who everyone is messaging or calling.

47:58We do not track the messages people are sending one another for ads. We do not sell user data. We approached Google, too, on Meredith's assessment of their business model, but we've had no response. And that AI Christmas thought? The conversation with Mustafa Suleiman of Microsoft is in the feed. Just scroll back to last December to find it. Finally, the team. The producers are Jessica Beck and Chris Martlu. The video producer is Andy Haywood. Social media is by Alex Morgan. audio mixing by Richard Ward. Our music is by Bart Warshaw and the executive producer is Louisa Lewis. At Bloomberg Weekend, thanks to Brendan Francis Newnham and our executive editor, Catherine Bell.

48:47And until next time, goodbye.

48:54As industries evolve faster than ever, companies need an environment that accelerates strategic growth and Michigan delivers on that promise. From emerging startups to global enterprises, Michigan offers what executives value most, a resilient, innovative ecosystem, diverse communities that attract top talent, and a quality of life that supports work-life balance. With our unified Team Michigan approach, businesses scale faster and compete at the highest level. Michigan, pure opportunity. Seize your opportunity at michiganbusiness.org. When you're running a business, the best days are the ones where priorities stay on track.

49:29For midsize and large companies, risk can affect multiple parts of the organization at once, from property and liability to cyber and regulatory challenges. At that level, managing risk becomes an ongoing discipline. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. And when losses do happen, that work is paired with insurance coverage shaped by years of underwriting, risk engineering, and claims experience. Learn more at thehartford.com slash risk mitigation. Policies provided by Hartford Fire Insurance Company and its property and casualty affiliates, Hartford, Connecticut.

50:05Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drinkware to tech and totes, it's easy to find the right fit for your brand and budget with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain.

From the publisher

Meredith Whittaker has spent years arguing that privacy is a prerequisite for a free society.

As the president of the nonprofit foundation behind the encrypted messaging app Signal, she now finds herself defending that principle in the face of mounting pressure from governments and tech companies alike.

In this conversation with Mishal Husain, Whittaker says she believes business models that rely on data collection, the rise of artificial intelligence assistants and even well-meaning efforts to protect children online risk undermining private communication. She also explains why Signal would rather leave a market than weaken the product’s encryption.

If you enjoyed this episode, listen to Mishal’s conversation with Mustafa Suleyman from Microsoft AI: Mustafa Suleyman Isn’t Like Everyone Else in Silicon Valley

Contact The Mishal Husain Show mishalshow@bloomberg.net

Subscribe today on Apple Podcasts and Spotify.

See omnystudio.com/listener for privacy information.

More from The Mishal Husain Show

All 53 episodes
AI Wants Your Life: Tech Boss Meredith Whittaker Says NoThe Mishal Husain Show · 45 min
Listen in VO