AI Everything Today

Overview: Tim Berners-Lee on the web, AI, and control

Sir Tim Berners-Lee, the inventor of the World Wide Web, spoke at length about the web’s origins and its future. He named the original slogan, “This is For Everyone,” and traced how the web moved from that idea to a world dominated by a few large platforms. He discussed projects he leads, including Solid and the company Inrupt, plus the influence of OpenAI and other AI firms on how people will access and use web content.

In the interview he covered a range of topics, including the successes and limits of the Semantic Web, the rise of AI-powered browsers like OpenAI Atlas, risks of AI extracting web data, and the potential of personal data wallets to restore user control. He mentioned specific players such as Cloudflare, Apple, and the Chromium and WebKit browser engines as examples of infrastructure power and market concentration.

Why this matters to everyday users

These conversations affect ordinary people because they shape how we find information, who controls our personal data, and how creators get paid. If AI tools pull content from across the web and present it directly to users, that can change advertising and subscription models that support news sites and independent creators.

Berners-Lee wants architectures that let people control their own data. He argues that better standards and tools could allow AI to use web data while preserving publishers’ value and user privacy. That has direct consequences for your privacy, the cost of online content, and the diversity of trusted sources available online.

Key ideas explained

1. Centralization versus openness

Berners-Lee warned about market concentration. Search, browsers, and social platforms have centralized power. That means a few companies can influence which content is visible, and what data is collected about users. Centralization reduces digital sovereignty for both individuals and publishers.

2. The Semantic Web, what worked and what did not

The Semantic Web was an effort to give web content clear, machine-readable meaning. Elements of that idea succeeded, such as linked open data and Schema.org, which helps search engines understand article headlines, dates, and ratings. But the broader vision did not fully take off. Generative AI now extracts useful structure from ordinary text, so in practice AI is doing part of what the Semantic Web aimed to do.

3. Agentic browsers and AI-powered interfaces

New products described as agentic browsers combine browsing with AI agents that can act on your behalf. Examples include experimental projects like OpenAI Atlas. These agents can fetch, summarize, and even transact using web services. That could simplify tasks, but it also creates risks if agents take value away from publishers or service providers.

4. Solid, Inrupt, and personal data wallets

Berners-Lee promotes Solid, a set of standards and tools that let people store personal data in their own “pods” or wallets. Inrupt is a company built around Solid to help bring these ideas to market. The goal is that apps and AI agents access personal data only with permission, and users keep ownership and control over that information.

5. The DoorDash problem and agentic commerce

Berners-Lee used an example similar to the problem restaurants face when delivery platforms take large commissions. If AI agents buy on users’ behalf, they might choose the cheapest or most convenient option, cutting out the seller’s brand and relationships. That disintermediation can reduce incentives for creators and service providers to produce quality work.

6. Local versus cloud AI

There is a trade-off between local AI running on your device, and cloud-based AI that runs on company servers. Local AI can be better for privacy and user control, while cloud AI often offers more compute power and convenience. Berners-Lee said the long-term ideal is powerful local inference, combined with interoperable data, so users can benefit from AI without losing control.

7. Standards, governance, and possible coordination

He emphasized the need for coordinated standards and governance for AI and the web. One suggestion is a collaborative effort similar to CERN, but for AI and web standards. The aim would be to ensure interoperability, protect users, and prevent a few companies from making unilateral technical decisions that shape the whole web.

8. Infrastructure power plays

Companies that operate core web infrastructure can set policies that affect crawling and content access. Cloudflare’s Content Signals Policy is an example where an intermediary can influence how crawlers and AI systems access content. When intermediaries assert control, it creates tension with the idea of an open web.

9. Browser engines and market concentration

Berners-Lee raised concerns about Chromium’s dominance as an engine, and the limited competition on mobile devices where WebKit is the main engine on Apple devices. Opening mobile platforms to more browser engine competition could spur new kinds of web experiences and reduce the power imbalance between platforms and publishers.

10. Monetization and the open web

AI-driven interfaces may bypass traditional ad or subscription revenue, especially when they present condensed answers rather than links. That could reduce income for news outlets and independent creators. Berners-Lee argues for new monetization models, including micropayments and standardized ways for AI agents to recognize and reward creators.

Practical steps Berners-Lee recommends

• Develop and adopt standards for personal data control, such as Solid pods and interoperable data wallets.
• Create incentive structures for AI agents to respect publishers and service providers, so producers are not disintermediated.
• Encourage browser engine competition and platform openness, especially on mobile devices.
• Explore coordinated governance for AI and web standards to ensure interoperability and public interest protections.
• Support research and tooling that make local AI inference more practical for users, balancing convenience and privacy.

What ordinary readers should watch for

• Changes in how search results and AI assistants cite or link to original sources.
• New products offering personal data wallets or Solid-compatible services.
• Policy moves by infrastructure providers that affect crawling, scraping, and content access.
• Shifts in browser and mobile platform rules that influence which engines can run on phones.
• New monetization options for publishers that try to work with AI agents, such as micropayments or standardized attribution signals.

Key Takeaways

• Tim Berners-Lee does not believe AI will destroy the web, but he sees real risks from centralization and unregulated agent behavior.
• Personal data wallets and Solid are proposed solutions to give users back control over their data.
• AI can and does extract semantic information from unstructured content, reducing the need for earlier Semantic Web approaches, yet standards still matter.
• Market forces alone may not protect openness and interoperability. Standards and governance could help preserve the web as a public resource.
• Practical changes to browsers, payment flows, and policies will shape whether the web remains open and equitable for creators and users.

Frequently Asked Questions

Does Berners-Lee think AI will end the web as we know it?

No. He said AI will not destroy the web, but AI can change how the web is used. The risk is that a few companies or designs could capture too much value and reduce choice for users and creators.

What is Solid and how would a personal data wallet work?

Solid is a set of standards that store personal data in separate user-controlled spaces called pods. A personal data wallet holds your profile, preferences, and private data. Apps and AI agents request access, and you decide what to share. The idea is to shift control back to the user.

Why should I care about agentic browsers?

Agentic browsers use AI to act for you, such as booking flights or summarizing research. They can save time, but they also decide which vendors and content you see. That affects the income creators receive and the diversity of information available.

Can regulation solve these problems?

Regulation can help, but Berners-Lee believes it must be paired with technical standards and voluntary cooperation. Rules alone may not produce interoperability or user-focused architectures without coordinated technical work.

Conclusion

Sir Tim Berners-Lee offered a measured view of the web’s next phase. He named the threats that come from concentration and unchecked agent behavior, and he proposed concrete alternatives such as Solid personal data wallets and stronger standards. For ordinary readers, the question is not whether AI will replace the web, but how we will design the systems that mediate our access to information, manage our personal data, and reward the people who produce online content.

How the industry responds, and whether standards and governance keep pace with AI-driven change, will determine whether the web remains open, diverse, and useful for everyone.