The digital landscape is shifting in fundamental ways. Where users once navigated the internet independently, artificial intelligence systems now increasingly perform that role on their behalf. Millions turn to tools such as ChatGPT and Claude for answers, product recommendations, news summaries and decision-making support. The next wave promises autonomous agents—including Meta's Muse—that will handle tasks like comparing insurance options, booking medical appointments, managing household supplies or rebooking flights without requiring users to visit a website at all.
The appeal is obvious. Yet this convenience masks a deeper concern: the concentration of power these intermediaries wield. AI systems now determine which information reaches users first, which products get recommended and which businesses gain visibility. The stakes are substantial. Bain & Company projects that by 2030, purchases "initiated, influenced, or completed" by AI agents could represent hundreds of billions in US e-commerce activity.
A troubling pattern is already emerging. Leading artificial intelligence companies have begun signing licensing agreements with major news publishers, arrangements that often ensure those outlets receive prominent placement in AI-generated responses. While such deals can provide publishers with compensation and increased visibility in an AI-mediated ecosystem, they also demonstrate how a small number of dominant platforms could become the primary intermediaries between consumers and the broader internet.
This dynamic mirrors historical precedents. Businesses may soon find themselves competing not just for customers but for favorable treatment within a handful of proprietary AI systems—a situation analogous to fighting for shelf space in physical retail or app store visibility on smartphones. If AI agents preferentially direct users toward companies that pay for preferred positioning, sign licensing deals or integrate directly with their platforms, the result could be a fundamentally closed internet.
How the Early Internet Avoided Gatekeeping
The internet's early decades largely escaped this kind of gatekeeping because of deliberate architectural choices. Shared technical standards prevented any single entity from controlling participation. Any business could register a domain name, build a website and reach users across the web without paying intermediaries for access. The Domain Name System, or DNS, created a universal naming convention that allowed browsers to connect users with legitimate websites without restricting which sites could be discovered or charging businesses for visibility.
This openness was not accidental. The internet's architects intentionally distributed power rather than concentrating it, creating conditions where small startups could compete on equal footing with established corporations.
Building Open Standards for AI Agents
An analogous approach could prevent AI from becoming the internet's new toll booth. An open Agent Name Service would function similarly to DNS for the current web. Rather than requiring businesses to negotiate access through a handful of dominant platforms, such a system would enable AI agents to discover and interact with businesses and services across an open network, independent of which platform operates those agents. AI systems could then interact across a more competitive and open ecosystem rather than operating within a few closed environments.
Efforts toward this goal are already underway. Google's Agent2Agent protocol, now managed by the Linux Foundation, enables AI agents from different companies to communicate with one another. Anthropic's Model Context Protocol, similarly housed under the Linux Foundation, establishes a standard method for AI systems to connect with external tools and data sources.
The underlying principle matters more than any specific technical implementation: the next phase of internet development should rest on open standards rather than proprietary gatekeeping mechanisms. As artificial intelligence reshapes how people engage with the digital world, society faces a critical choice. The internet can remain an open marketplace where competition drives innovation and accessibility, or it can become a system where a few platforms extract tolls from everyone seeking to reach users online.



