Is AI Mode killing the internet’s search-based business model?

When AI chatbots burst onto the scene, there was chatter that they would kill Google search and the surrounding business models. 

The latest fiscal results are shedding some light on this prediction, and it seems that it is indeed affecting publishers and companies depending on Google search traffic. 

Google, in a bid to stay ahead of chatbots like ChatGPT and Claude, launched its own AI mode. 

However, that is creating another problem, as more users use AI mode instead of visiting the websites providing the information. 

There was a simple business model. Google crawled and indexed websites and displayed those links to users and directed traffic towards publishers. 

Those visits provided advertising revenue, attracted subscribers for higher-value information and allowed websites to build audiences. 

Artificial intelligence is challenging this model. 

Google’s AI Overviews and AI Mode provide users with information directly on search-result pages, reducing the need to click through to individual websites. 

The consequences are beginning to appear in company earnings.

Reddit reported a sharp decline in search referrals.

Freelance platform Fiverr also highlighted the impact of AI-related disruption on traditional search engine optimisation (SEO) work. 

Google is also facing challenges in generating advertising revenue from AI mode compared to revenue from conventional search.

The development raises a broader question about the sustainability of the open internet. 

If websites lose the traffic that finances their content, what happens to the information that AI models themselves depend on?

Industry executives believe the shift could represent a lasting change in the economics of the internet, with implications for publishers, companies and users. 

Google’s AI transformation is changing the economics of search

Google started AI Mode as an answer to ChatGPT.

ChatGPT’s introduction made companies question whether conversational AI would eventually replace a search engine.

The initiatives appear to be generating greater engagement. 

Google reported its highest query growth in five years. 

However, increased search activity does not necessarily translate into higher traffic for websites or greater advertising revenue for Google.

Alphabet’s third-quarter results highlighted this. 

The company reported search growth of 12%, down from 19% at the beginning of the year, while search and advertising revenue fell short of Wall Street expectations.

Google also warned of further pressure on search-revenue growth in the fourth quarter.

Another indication of the changing economics is the deceleration in paid-click growth. 

Google’s paid clicks increased 13% year over year in the first quarter, but the growth rate slowed to 8% in the third quarter.

AI Mode could be contributing to this shift.

Unlike conventional search, where users typically encounter several links and advertisements before visiting a website, AI Mode attempts to provide a more comprehensive response within Google’s own interface.

This creates a different user experience. Consumers can ask follow-up questions, compare information, and explore topics without necessarily leaving Google.

For the search giant, the challenge is converting that engagement into revenue. 

For publishers, the problem is users tend to stay with Google’s AI Mode and end up not visiting their sites. 

Philipp Schindler, Google’s chief business officer, acknowledged during the company’s earnings call that AI Mode currently generates lower monetisation than traditional search, despite recording more queries.

Google is attempting to address the issue by expanding advertisements in AI Mode to additional countries and experimenting with new advertising placements as conversations progress.

Reddit’s traffic decline offers an early warning

The consequences of this changing relationship are becoming particularly visible among websites that have historically depended on Google for audience acquisition.

Reddit is one of the clearest examples.

The platform’s shares plunged approximately 29% on October 30 after its third-quarter results and outlook disappointed investors.

Investors were particularly concerned about weaker-than-expected user growth and the deterioration in search referrals.

Reddit CEO Steve Huffman said search traffic had declined by more than one million referrals per week during October.

He also warned that daily active users coming from search could decline by more than 10% in the fourth quarter.

Huffman suggested that Google’s AI Overviews were becoming a more prominent part of the search experience, pushing traditional website links further down the results page.

“What started as a feature is now the full search experience,” he said during the earnings call. “And, at a minimum, it pushes everyone down the search page. And then, on top of that, the experiences themselves are not traffic drivers.”

He added that users continued to want Reddit’s content, but Google was no longer directing visitors to the platform at the same rate.

“Google is crawling us as hard as ever, so the demand hasn’t changed from their point of view either. The reality is, the AI mode does not drive traffic, not for us and not to anybody,” Huffman said on the post-earnings call.

The distinction is important because Reddit’s business model benefits from user engagement, conversations, and the extensive archive of discussions available on its platform.

Historically, Google helped introduce new users to those discussions. 

AI-generated answers can now provide information extracted from similar conversations without requiring users to visit Reddit.

The platform is responding by tightening access to its content. 

Reddit said it would begin trialling a requirement for users to log in to access the website in international markets before the end of the year.

The move is intended partly to restrict AI bots from scraping its content.

Publishers face a weakening incentive to produce free content

The problem extends beyond Reddit.

Fiverr CEO Micha Kaufman described the changes as an earthquake across the internet, arguing that organic clicks were declining as more consumers turned to AI assistants to find information.

“The internet as we’ve known it for the past 20-plus years is going away, and it’s going to completely change,” Kaufman said. “It’s dying.”

Fiverr’s third-quarter revenue declined 15.8% from the year-earlier period.

In its shareholder letter, the company highlighted the impact of AI-related disruption on traditional SEO services, saying the changes had affected both low-value and high-end work on its platform.

The development could have broader consequences for digital publishing.

Many online businesses depend on organic search traffic to attract readers and generate advertising impressions. 

Others use search referrals to acquire customers, promote subscriptions or direct visitors towards commercial products.

When AI-generated summaries answer questions directly, fewer users may reach the original websites.

Monica Tartau, Head of Marketing at Aurora Labs, believes this represents a structural rather than temporary change.

“It looks more permanent than temporary to me,” Tartau told Invezz.

She argued that the traditional relationship between Google and publishers is changing because the search engine increasingly answers questions itself rather than directing users to the websites providing the underlying information.

According to Tartau, this weakens the financial incentive for websites to continue producing freely accessible information.

“A lot of sites made money or justified the work because people clicked through from Google,” she said.

The expert added that when AI mode provides the answer before the user clicks, “that page view, ad, signup, or sale often doesn’t happen.”

The problem is particularly relevant for businesses that rely on large volumes of relatively inexpensive content and organic search traffic.

AI search could undermine the information it depends on

The potential consequences extend beyond the immediate financial pressure on publishers.

AI search engines rely on information produced across the internet. 

News organisations, specialist publications, research platforms and independent websites contribute material that helps AI systems answer questions.

If those publishers lose the revenue that finances their work, the supply of original information could deteriorate.

Michael Heinrich, CEO of 0G, described this as one of the more serious long-term risks associated with AI-powered search.

“Producing original information costs money,” Heinrich told Invezz.

. 

Heinrich expects publishers to respond in different ways. 

“Larger publishers can negotiate licensing agreements, specialist platforms may move behind paywalls, and smaller independent sites could simply publish less,” the expert noted.

That creates a potential feedback loop.

“The problem for AI companies is that they still need a steady supply of accurate, current information. If fewer people are producing original material, AI systems risk drawing from an increasingly repetitive pool of secondary content, including material generated by other AI systems.”

This creates a contradiction for the AI industry. 

The technology could become increasingly effective at extracting and presenting information while simultaneously weakening the financial incentives that encourage people to produce that information in the first place.

Can Google make AI Mode as profitable as traditional search?

If Google backs AI Mode, it needs to find a way to transfer the advertising economics from search to AI Mode. 

In traditional search, a user searching for a particular product or a service can see sponsored links alongside organic results. 

Advertisers can compete for those placements, while Google earns revenue from commercial interactions.

AI Mode offers a different opportunity.

Because users can provide more context and ask follow-up questions, the platform may gain a better understanding of their needs and purchasing intentions.

Heinrich believes this could eventually create new advertising opportunities.

“If a user asks an AI assistant to compare laptops for a particular budget, for example, there’s a very clear commercial intent that advertisers would be willing to pay to reach,” he said.

Google has already begun expanding advertisements within AI Mode, including to additional countries and through experiments involving advertisements that appear as conversations continue.

However, integrating advertising into AI-generated answers introduces another challenge: maintaining user trust.

In conventional search, sponsored links are generally distinguishable from organic results. 

Within a conversational AI response, the distinction between an objective recommendation and a paid placement could become less obvious.

Heinrich expects advertising decisions to continue reflecting user intent, relevance, conversion potential and advertiser bids.

But he cautioned that commercial incentives should not determine what an AI system presents as factual information.

If consumers begin questioning whether recommendations are genuinely useful or simply favour advertisers, Google could undermine the credibility of the product it is trying to monetise.

Tartau also noted that Google is already expanding advertisements in AI Mode, but said the company has acknowledged that these placements currently generate less revenue than conventional search advertisements.

She said Google’s expectation is that users conducting early-stage research may become more commercially valuable as they continue asking follow-up questions.

For now, however, the ability to generate more interactions has not necessarily translated into equivalent monetisation.

Are paywalls and login requirements a good solution?

As search traffic becomes less reliable, websites are likely to reconsider how they distribute their content.

One option is restricting access through subscriptions, registration requirements, or other forms of login barriers.

Reddit is already planning this. By limiting access, the platform will make it more difficult for AI systems to scrape its discussions. 

Tartau said that for Reddit, “a login stops some of that, gives them a direct relationship with users, and more control.”

However, login requirements are not necessarily an effective solution for every publisher.

Tartau said such restrictions could also prevent casual visitors from discovering a website through search.

For Reddit, the strategy may make sense because its value lies partly in its extensive archive of conversations and user-generated discussions.

For a conventional publisher, the calculation is different.

“But if your business depends heavily on people discovering you through Google, putting everything behind a login can be quite damaging. You’re effectively asking people to commit to your platform before they’ve had an opportunity to understand its value”, Heinrich added.

Experts expect many websites to adopt a middle-ground approach.

Michael Heinrich said there would be a hybrid approach where “publishers keep enough content publicly accessible to attract new readers while protecting premium research, proprietary data or specialist analysis.”

He believes publishers with differentiated content and established audiences could have greater opportunities to negotiate licensing agreements with AI companies.

However, smaller websites that depend heavily on search discovery could face greater difficulties.

The expert also highlighted an important distinction: restricting access through a login wall does not automatically establish ownership rights or prevent every form of AI use.

He added that a publisher’s negotiating position will depend on the uniqueness of its information, licensing arrangements and the technical controls used to enforce access.

“Most sites will probably do something in the middle: leave some stuff open so they can still be found, put the better material or tools behind a login, and try to build more direct relationships through newsletters and apps so they’re less dependent on Google,” Monica Tartau added. 

The internet’s next business model remains uncertain

Despite the hybrid model’s possibility, there is still uncertainty for the publishers. 

Funding the content that populates the internet is always a challenge. Free information was supported by advertisements driven through clicks. 

As AI models eat up advertisements and clicks, it threatens to close off the internet and make users pay for basic access. 

The internet’s search-based business model is therefore facing more than a decline in clicks. 

It is confronting a fundamental question about who gets paid when information is created, indexed, extracted, and delivered through artificial intelligence.

AI disclosure: AI tools assisted with research and drafting of this piece. An Invezz editor reviewed the facts, sources and wording before publication. Any opinions or analysis are Invezz’s, not those of an automated system.

The post Is AI Mode killing the internet’s search-based business model? appeared first on Invezz