The Achilles Heel in AI Search Dominance

Reddit's October 29 announcement that it will begin testing a primarily logged-in experience in selected markets pulls back the curtain on an issue that's been just below the surface for months. It's not just a change to Reddit's user experience. It's a warning flag about a fundamental vulnerability in the economics of AI search.

The move is understandable. Google search referrals to Reddit declined by more than one million per week on average during October, according to Reuters. Reddit CEO Steve Huffman attributed the decline, in part, to Google's AI search, which answers questions using information from websites without necessarily sending users to those sources.

The irony is hard to miss. AI search depends on information created by others, yet its success can undermine their business models and create a disincentive to share that information.

At the information crossroads

Content owners essentially have two choices. As Reddit proposes, they can restrict access to their information, putting content behind logins, subscriptions, or technical barriers. I expect we'll soon see another path: negotiating licensing agreements with AI providers and getting paid for access and usage.

Neither path is particularly comfortable for AI search providers.

If more authoritative sources restrict access, AI search risks losing diversity, depth, and credibility. The remaining accessible content could increasingly be dominated by organizations with commercial incentives to publish, promote, or influence AI-generated answers.

Licensing offers an alternative, but it introduces its own challenges. Large publishers will have the leverage to negotiate meaningful agreements. Smaller publishers, independent experts, and specialized information providers will find it more challenging. We may even see the formation of aggregators that package and license content from smaller providers.

The Reuters Institute's Journalism, Media, and Technology Trends and Predictions 2026 highlights the concern. Publishers surveyed expect search engine traffic to decline by an average of 43% over the next three years. It’s clear the respondents underestimated the risks based on the pace of change we're already seeing.

The same study also shows that only 20% expect AI licensing to become a substantial revenue source, while another 20% do not expect to generate any revenue from AI agreements. The Reddit announcement should be a wake-up call.

Meanwhile, Google is moving quickly to monetize the new environment. Its October 28 announcement expanding advertising in AI Mode included an interesting statistic: AI Mode queries are, on average, three times longer than traditional searches. Google is also introducing advertising into follow-up questions, extending commercial opportunities into ongoing AI conversations.

The search experience is changing. The economics supporting its information sources are still being worked out.

The next data frontier

The AI revolution is being fueled by access to the large and diverse datasets available online. Much of this data is digital, published, documented, and relatively accessible. But an important category remains largely untapped: the data generated by people, equipment, and processes operating in the physical world.

Consider manufacturing facilities, energy operations, logistics networks, and field service organizations. Every inspection, equipment interaction, quality check, maintenance activity, and safety observation generates valuable operational information. Much of it remains trapped in paper processes, disconnected systems, or the institutional knowledge of experienced frontline workers.

This is a different kind of data challenge. It is not simply about gaining access to existing information. It is about capturing information that has never been systematically digitized. Much of this data will be proprietary and valuable to the organizations generating it. But over time, we could also see the development of shared or licensable data assets that can be used to train, inform, and improve AI applications.

Platforms such as ROO.AI are addressing this gap by bringing AI into frontline workflows, guiding work execution while capturing operational data in real time. This creates a valuable two-way relationship: AI helps people perform work more effectively, while the work itself generates new data and intelligence.

Ultimately, AI value will not be determined solely by who builds the largest models or the most sophisticated search experiences. It will also depend on access to the highest-quality information, and how that information is verified, and the ability to continuously generate new data.

The next competitive advantage in AI for your organization may come from information that doesn't yet exist in the digital world.