Web 4.0 Has Been Declared. Here’s What Publishers Need to Know.
A 2026 manifesto from Sigil Wen and the Conway group introduced the idea of Web 4.0. It highlights changes already reshaping how publishers structure content, rights and data.
Earlier this year, Thiel Fellow Sigil Wen and an AI research group called Conway published a document describing what they term as Web 4.0. This is not a new internet standard and no global protocol has changed. What it is, however, is a thesis arguing that AI systems will increasingly act as economic participants online.
Wen announced the idea publicly on X alongside the manifesto:
What Web 4.0 actually is
Wen’s argument is that software will move beyond assisting humans and begin operating on their behalf in markets. That means AI systems accessing services directly, negotiating transactions and making decisions continuously rather than only when prompted.
The reason this matters to publishers is not because a new “version of the web” has formally begun per se. It matters because the behaviour being described is already emerging inside media businesses.
Where this affects publishers
Start with newsroom workflow.
Many publishers now use AI to help draft headlines or summarise documents. The more significant shift is that AI tools are being embedded inside production systems rather than used as separate chat interfaces.
Take Dalet as an example. Dalet is a media tech company that provides workflow software to broadcasters and large publishers. Its latest AI layer connects several stages of the production process. When raw video footage is uploaded, the system automatically creates a transcript, suggests metadata tags, identifies key moments and routes the content through approval workflows. Editors still make decisions, but the repetitive handling of files between systems is reduced.
When content moves through consistent workflows with reliable metadata, it becomes easier to search, package, license and measure. This type of infrastructure optimisation directly affects publisher economics.
Archives provide another example.
TIME last year launched a conversational interface that allows users to query its archive directly. Instead of searching manually through decades of reporting, users can ask questions and receive contextualised summaries drawn from structured material. For the publisher, this turns the archive from a storage system into a continuously usable content asset that drives engagement.
Licensing is also becoming more technical.
In the US, the Associated Press has defined exactly how its content can be used in AI deals, while in Europe publishers such as Axel Springer and the Financial Times have signed agreements that specify usage and compensation.
These deals rely on tracking how content is used, which means payments are increasingly based on measurable activity rather than broad claims about value. This is already visible in practice, as AI tools influence how licensing agreements are structured.
Practical Implications
For publishers, the implications are practical:
• Poorly tagged archives surface less often in AI-driven search and summarisation products.
• Unclear rights documentation weakens leverage in structured AI licensing negotiations.
• Inconsistent or unverifiable audience data reduces allocation in automated ad buying systems.
• Content that cannot be accessed cleanly through APIs is harder to integrate into distribution partnerships.
• Weak metadata makes packaging, syndication and measurement more expensive and less scalable.
“Web 4.0” is not an official new phase of the internet, but the changes it describes are already affecting how publishers’ content is discovered, licensed and monetised.
Publishers should review how their products are structured and exposed, starting with clean interfaces, consistent product definitions and reliable payment flows. If AI tools begin to compare or manage subscriptions, for example, those systems will rely on clarity rather than persuasion, which means offers, eligibility rules and pricing need to be explicit and defensible.
The near-term question now is whether your content, rights and audience data are organised well enough to compete in systems that increasingly determine what gets distributed, licensed and funded.
This shift is already happening.
what tools do you see / know that publishers are using to help with this? In my own work / research, I found that some are doing well but there are a lot of stake oil sales people selling to publishers...