WordPress Developers Reject New AI Feature for Core

A new proposal aims to embed a Knowledge Custom Post Type (CPT) directly into WordPress Core, but the idea is already facing significant pushback from developers. The feature, which has already been introduced as an experimental addition in Gutenberg, is designed to act as a central hub for storing website guidelines,such as brand voice, tone, image preferences, and editorial rules. Its stated purpose is to serve both human users, like editors and contributors, and internal AI agents and tools. However, the response from the developer community has been largely negative, with many questioning whether this feature addresses real-world needs.

To understand the proposal, it helps to know what a Custom Post Type is. Currently, WordPress ships with two default post types: Posts and Pages. Custom Post Types, often added via plugins, allow site owners to create tailored content structures. WooCommerce, for example, uses a CPT called “product” to help shopkeepers manage inventory.

The Knowledge Custom Post Type was first proposed in February 2026 and was integrated into Gutenberg as an experimental feature a month later. It supports several knowledge categories: Site (goals, personality, audience), Copy (tone, voice, vocabulary), Images (preferred styles, colors, moods), Blocks (per-block rules, like short sentences for paragraphs), and Additional (accessibility, linking practices, formatting rules). The idea is that this centralized repository would be accessible to everyone working on a site, including humans, AI agents, tools, and plugins.

But confusion has emerged over who the feature is really for. While the public proposal pitches it as a tool for both humans and AI, the GitHub repository for the feature tells a different story. It describes the Guidelines CPT as storing “site-wide editorial rules” and notes that “as AI-powered tools integrate with WordPress, a recurring need is emerging for sites to store different kinds of persistent, structured knowledge that shapes how agents interact with the site.” There is no mention of human utility on that page. This discrepancy between the public description and the technical specification has led many to question whether the WordPress core team truly understands the feature’s purpose or its intended audience.

This is just one of at least six reasons developers are pushing back. The proposal has sparked heated debate in the private Dynamic WordPress Facebook group, where 29 comments overwhelmingly opposed merging the feature into core. The main objections include: the feature should remain a plugin, not be baked into core; it adds unnecessary bloat; there are higher priorities, with native multilingual support cited as the most critical missing functionality; doubts about whether the proposal was fully thought through; the sense that this is an AI-focused feature, not a user-facing one; and the suspicion that it primarily benefits Automattic’s enterprise users on WordPress.com, rather than the average site owner.

Resistance is not confined to social media. On the official WordPress announcement, commenters raised similar concerns. User mrwweb wrote: “I know it says this feature is provided for both ‘author-facing and agent-facing’ applications, but it feels like AI/LLMs are driving the conception of the feature. Further, the underlying assumption that ‘Most sites already have content standards’ does not strike me as accurate. That is a very real need and use-case, but I’m not sure it’s one that would justify a new core feature on its own.” Namith Jawahar called it “an unnecessary overreach,” adding that developers should have the freedom to decide for individual sites. Aaron Jorbin acknowledged the feature’s potential but said it “feels incomplete” and needs more time before being added to version 7.1.

This is the second time in the WordPress 7 release cycle that a new feature has faced serious scrutiny. Earlier, version 7.0 saw developers question the value of Real-Time Collaboration (RTC) , which was supposed to be a hallmark of Phase 3 collaboration. The pattern suggests a growing tension between the platform’s direction and what the community believes is truly needed.