BRUCE SCHNEIER - AI companies gear up to sell ads

Dr Bruce Schneier - Lecturer in Public Policy, Harvard Kennedy School

INTRO: Are you ready for paid advertising in your ChatGPT? To recuperate the rising costs of Artificial Intelligence, it seems Big Tech is now poised to introduce advertising into your chatbot – whether you want it or not. RTRFM’s Allan Boyd spoke to a cyber-security guru to find out more…

Conversation Article: Could ChatGPT convince you to buy something? Threat of manipulation looms as AI companies gear up to sell ads -  January 15, 2026 

Transcript

SPEAKER_02 0:00

Whether we like it or not, and whether we know it or not, it has been a rapid embedding of AI in our data live. But could chat GPT to manipulate you into buying stuff? Despite being earlier applied to it and following the footsteps of social media to monetize consumer attention. So will we see AI-driven advertisements inside chatbot competition? Is this the identification of chat GPT as yet another online platform becomes less unusable? To talk about this and the persuasive nature of generative AI chatbots, I'm joined by cybersecurity guru, Dr. Bruce Schneier. So why is advertising in chat GPT an issue?

SPEAKER_00 0:46

So it's an issue because the uh AI companies are at this point pretty desperate to figure out ways to make money. These are new companies are losing incredible amounts of money. And they're floating the idea of adding advertising. Uh Google has talked about it in their model. Recently, OpenAI talked about adding advertising. We don't know what that's going to look like, but the worry is that it will be highly persuasive and hard to spot. Because uh, unlike a Google search, a chatbot is conversational. So if you're talking to a chat bot and I don't know about a vacation and it suggests a particular hotel chain or destination, you don't know if they're getting a kickback off the back end. And that's the fear that it'll be deceptive and manipulative. And because it is an agent, because it is someone who talks to you in natural language, then that is just much more dangerous, potentially.

SPEAKER_02 1:36

Because of its persuasive nature?

SPEAKER_00 1:38

So we don't know how persuasive they're going to be, but potentially they could be more persuasive because they speak in a natural language, because you are likely to think of them as a friend. They're gonna act friendly, they'll act authoritative, they're obsequious, they uh they agree with you, they you know, they they're designed for you to respond to. You know, all these things peg our human reciprocal tendencies to think of this as another person and not want to offend it. So it really does play on a lot of our our human frailties, unlike even your Facebook feed or uh you know Google search. You know that a particular Facebook post is sponsored. You see that. You can see what a sponsored result looks like in Google. How do you know if it's sponsored when it's a conversation?

SPEAKER_02 2:28

Yeah, unless they unless they're telling you it's embedded inside the interface. And they might.

SPEAKER_00 2:31

So we we don't know we don't know. Um we do what we don't know what the interface will look like. And you can imagine that they'll be straightforward and tell you, but that's not the way to bet. You know, these companies want to make money, and they'll make money by advertising that's effective, not that's honest.

SPEAKER_02 2:45

Yes, and it's very different to how Google works in that you ask it a question and it gives you a website to go to. This is far more persuasive.

SPEAKER_00 2:52

Right. It's more interactive. It's more interactive, and you think of it l more as a friend and less as a service.

SPEAKER_02 2:58

As you say in the article, it's like having a conversation with an author rather than just spitting out a website.

SPEAKER_00 3:04

And that's a good thing, right? I mean, so a lot of work is being done on uh AI as an interactive textbook, right? So instead of reading a dry textbook, a chapter, you engage with a conversation with an AI. So it has incredible learning potential. But when you start adding commercialization into the mix, it just it becomes a different animal.

SPEAKER_02 3:23

The uh the other point that AI is fundamentally untrustworthy, and so that makes it also problematic.

SPEAKER_00 3:29

Well, and and this is a problem with the technology, it's two problems. Problem of the technology, so we know that it makes mistakes, and that's okay. Humans make mistakes, we're good at the things that make mistakes. But also it is being run by a for-profit corporation, right? An American for-profit corporation. Yes. A bunch of white male tech billionaires who are producing these services which you're using for free for their benefit. They're not doing it out of charity. So I mean by untrustworthy, I mean that they're not on your side. It's really a good way to think about it. They have some hidden agenda. And advertising is a very powerful hidden agenda.

SPEAKER_02 4:03

Surveillance capitalism is another uh issue, I suppose. That's it.

SPEAKER_00 4:07

Well, this is an example of surveillance capitalism. So presumably, the advertising that ChatGP does is not gonna be random. It's not gonna in the middle of a conversation spout, you know, drink this soft drink, buy this brand of car. It will be embedded in the conversation where it naturally flows, which means the system has to spy on you to know what it is you're talking about and thinking about, so it knows when to suggest a car or a restaurant or a vacation destination. Right? And it's so it's in context. So this is, in a sense, the next chapter of surveillance capitalism.

SPEAKER_02 4:41

And because Google does collect so much data, it can even ask you what color car you might want to buy.

SPEAKER_00 4:46

We don't know. I mean, the the fear is of course that it'll be much more persuasive because it knows much more about you and how to influence you. You know, some of that is uh is not real. I mean, I think some of that is is uh the company's marketing want to convince you want to convince you that its advertising is also powerful. But I think some of it's real as well. And I think AI will make it more real.

SPEAKER_02 5:05

Aaron Powell So we just don't know what the business models are really, do we?

SPEAKER_00 5:08

You know, we we don't. I mean we do know that there is not a business model today, that these companies are widely unprofitable, that a lot of the uses in in corporations are not panning out. They're looking uh for a consumer model. Agents is is a model they're thinking of, but the question is will people pay for an agent? They don't know. Advertising is is a model by which you can give the service to consumers for free and make money off the back end. So it's really powerful, but it also is scary.

SPEAKER_02 5:35

It sounds quite scary. So what can we do? What should we do now that we know it's coming?

SPEAKER_00 5:38

You know, this is a matter of policy. I mean, I mean I think we've learned a lot from our inability to regulate social media. Yes. Uh which in a lot of cases we've just let run ambuck. I mean, Australia is a is a recent exception to that, uh with with uh regulations on accounts of children. But that's a sort of one thing of many that needs regulating. Um I think it's government has to step in and regulate the industry because it is so potentially coercive.