What is WebAI?
WebAI studies how artificial intelligence changes the Web across connected dimensions: content production, documentary discovery, agent interactions, structured knowledge and evaluation of results. The name refers to an educational research project, not an official protocol replacing the existing Web.
Several shifts within one Web
A resource may be AI-assisted in its writing, published as HTML, described in JSON-LD, indexed by search, summarized by a generative interface and visited by an agent. These mechanisms differ and complement one another. The source remains a standard Web resource and each transformation needs its own evaluation.
From production to decisions
Generating a text or image is not the same as retrieving evidence or performing an action. For each system, document the model, data, references, permissions and human oversight. This distinction protects users from the assumption that fluent answers or automated interfaces carry scientific authority.
Web foundations remain essential
HTTP, URLs, HTML, CSS, JavaScript, accessibility and open formats remain the foundation of WebAI experiences. Systems adding AI search or action capabilities do not make semantic structures obsolete. Organized, readable and referenced pages benefit both people and software systems.
Example: a learning encyclopedia
An encyclopedia can publish dated dossiers, connect important claims to primary documents, support corpus search and offer labs that simulate a method without pretending to call a live model. An agent should not publish or edit a page before rights checks and accountable approval.
Four guiding WebAI questions
Ask: who produced the information? How was it discovered? Who is authorized to act on it? What evidence supports reuse or publication? These questions separate technical possibility, local prototypes, deployed systems and verified claims.
Explore domains without conflating them
WebAI research may involve linked data, RAG, local inference, agents, provenance or accessibility. No domain solves every other problem. The portal compares approaches, maturity and limitations instead of suggesting one inevitable future.
Core steps
- 01Define the question and audience
- 02Separate creation, retrieval, action and evidence
- 03Connect original standards and sources
- 04Test a concrete scenario
- 05Examine outcomes and permissions
- 06Publish a dated summary
Three questions to ask
Is WebAI an officially adopted standard?
Must generated pages still follow Web standards?
Does a citation alone prove the truth?
Who authorizes an agent action?
Read the source documents
Versions, reuse terms and document status can change. Check them on the original sites.