EmTechMIT, Cambridge, MA: Brian O’Kelley, a pivotal figure in digital advertising since the early 2000s, opened his talk by observing that the very fabric of internet advertising is undergoing a seismic transformation as AI accelerates the death of traditional web channels. “Advertising is not going away anytime soon but the Web is already on its way out as a way of reaching customers. AI is only going to accelerate that shift”.

Reflecting on two decades of change, O’Kelley explained, “In 2005, I invented the first advertising exchange, the highest-paying advertiser gets to show you an ad. That happens about a trillion times a day.” This moment sparked a technological revolution that enabled dynamic, per-user advertising and established the underpinnings of today’s giant digital ad economy.

The New Operating System: AI as Interface

The “death of the web” isn’t about fewer websites, it’s about user attention and interaction moving from open web browsers to apps, closed platforms, and now, AI-driven conversational interfaces. “Most people of all ages are getting most of their content from YouTube [and] social media really not from the Web. The death of the Web has been coming for a really long time and now it’s just happening really fast”.

O’Kelley declared, “The new operating system of the internet is going to be ChatGPT.” Where Amazon today dictates commerce journeys, generative AI platforms promise to be the go-to gateway for everything, shopping, research, even advice. “My primary interface to so many things is going to come through this new interface and I say it’s an operating system because ChatGPT just announced apps”.

This means “all these things that we think of as apps on our phone are going to be apps in our chatbot,” whether it’s Gmail, Yelp, or Booking.com. The battle to be a user’s default “app” in this interface will rival the browser wars and mobile platform battles.

The Blurring Line: Content vs. Advertising

A core tension of trust emerges at the blurry edge of advertising and organic content. “How will we differentiate between these two things?” Brian asked. For now, platforms like ChatGPT do not insert overt promotional content, but the pressure is coming fast: “All the major players announced deals with, for instance, Walmart, that’s going to become immediately a place where all these companies are going to throw immense amounts of money”.

He predicted that this new era of sponsored content will move far faster than previous transitions: “How long do you think this is going to last? I think we have months. Honestly.” The economic force is simple: “OpenAI is going to lose $27 billion this year. So they probably can’t get that from consumers and subscriptions. So, I think that’s going to be happening very quickly”.

Algorithmic Nudges and Ethical Hazards

Rather than separate ads from content, AI could blur them indistinguishably: “The line between ads and sponsored links is very hard to distinguish even on Amazon. When you’re listening to a podcast and the host switches from talking about football to talking about the upcoming trip to Tennessee, but you can now take a go on, it’s a very thin line”.

Brian detailed the coming risk: algorithmic guidance may not identify ads, but could nudge user outcomes invisibly: “It will get more pernicious. There’s never an ad, it just nudges the content in ways that we can’t see. So, it’s an algorithmic bias, not an explicit [ad]”.

He joked, “Wouldn’t it be wild if I wanted to say my company’s name more than my competitor’s name so anytime, it’s about to randomly choose tokens. Scope3. I’ll pay you a dollar every time you mention Scope3, isn’t that cool? (or terrifying)”.

Trust and Transparency in the Agentic Age

One audience member worried: won’t this shift erode trust if AI chatbots always push Walmart links, for instance? O’Kelley responded, trust will hinge on transparency and choice, but predicted the options won’t be easy or durable. “I guarantee you will compromise the consumer experience, and it will compromise the quality of the output in one sense. But I don’t think they’ll do it in a deceptive way. I think they’ll do it in a very transparent way that you can turn off, just like you can turn off ads on Meta or not be tracked. You just have to click a lot of times. It’s really hard to get it to stick because they’ll just change the policies”.

He posed a sobering dilemma: “Advertising has funded the development of AI. Google is ad funded, Meta is ad funded, Amazon is [ad funded]… advertising is very deeply tied into the AI economy”.

Generative Engine Optimization, The New SEO

To stay relevant, brands must shift from web SEO to what O’Kelley called “Generative Engine Optimization” (GEO), reverse-engineering what info large generative platforms surface in responses: “There’s a lot of companies now looking at how ChatGPT chooses what information to share when you ask it a question by reverse engineering the algorithm. It will help you change your website to get better results”.

“It’s what search engine optimization was for Google. So much of your traffic is coming from either AI chatbots or from AI summaries on Google, you have to optimize traffic that you don’t ever see”.

Data, Ownership and Individual Agency

I asked about personal ownership of data. Despite academic optimism, O’Kelley was skeptical about people taking ownership of their data. “I’ve heard for 30 years about humans owning our own data. I am a self-proclaimed expert, and yet for some reason when I walk into CVS and it asks me for my phone number, I type it in. I literally don’t know what you get for typing in your phone number”.

He lamented, “Trust in big tech is like zero and yet we all use the engines and search tools every single day. I think we happily give away our data constantly for, you know, small potatoes, and yet the internet economy is working incredibly well in a world that you could pay nothing and have access to the world’s information at your fingertips”.

The Human Edge and the Future of Work

One concern is how AI might “train” on human-generated data and automate away creative work. O’Kelley noted, “Their challenge is that even though they know a ton about me, the value I provide is not just the textual corpus, it’s also relationships, physical presence, the ability to collaborate and learn and actually have some kind of lived experience in the world. The people whose jobs aren’t as inherently creative as mine, I think a lot of jobs really look a lot like that, right?”

He cautioned that the pace of disruption is increasing: “This is just gonna happen faster and faster. I’m doing amazing things in my spare time because I have these agents to do it for me. How many people can harness this?”

Final Reflections

Brian O’Kelley’s session left the technologist community with pressing questions:

• How do we safeguard trust as the boundary between content and advertising blurs?
• Can transparency and user choice realistically counteract economic incentives for subtle or algorithmic manipulation?
• Will consumers demand and enforce data ownership, or will convenience and price keep them locked into the data-for-services bargain?

He summed up the paradox: “Advertising is not going away; it’s evolving. What’s at stake is not just commerce, but the very architecture of digital trust in the AI era”.

For more information, please visit the following:

Website: https://www.josephraczynski.com/

Blog: https://JTConsultingMedia.com/

Podcast: https://techsnippetstoday.buzzsprout.com

LinkedIn: https://www.linkedin.com/in/joerazz/

X: https://x.com/joerazz