Why Proposal To Label Sections Of AI Generated Content Is Controvesial

A new proposal was published for creating an HTML attribute that can be helpful for notifying crawlers what part of a web page is generated by AI. The proposal is quickly becoming relevant because of new rules coming into effect in Europe this summer, but some are questioning whether this is the right solution to that problem.

What The AI Disclosure Proposal Is

An explainer for the new proposal says that the problem this solves is to there is currently no way to indicate that a section of a web page is generated with AI. What they’re proposing is a page-level meta tag that communicates a uniform degree of AI generated content on the page and also a section-level disclosure for when just a section of a page is AI-generated.

Page-Level Disclosure

The page-level disclosure uses a meta tag with a choice of attributes that communicate how much of the content is or is not AI-generated.

The choice of attributes are:

• “ai-generated” (AI-generated with human prompting and/or review)
• “ai-assisted” (Human-authored, AI edited or refined)
• “autonomous” (AI-generated without human oversight)
• “mixed” (Means that sections of the page have different levels of AI involvement)
• “none” (No AI involvement)

Element-Level Disclosure

The Element-Level disclosure relies on an attribute that can be used on any HTML element. The proposal uses two examples that rely on Semantic HTML, which can be considered to kind of exceed the parameters for which the elements were created (more on that later).

The proposal uses an example of the

semantic HTML element:

Six-Month Investigation: City Budget Shortfall

Our reporters spent six months reviewing financial records…

The proposal also uses an example of the

AI Summary

The investigation found a $4.2M discrepancy in the city’s
infrastructure fund, attributed to misclassified expenditures…

AI Disclosure

The proposal was created by David E. Weekly (LinkedIn profile), who noted that there are currently proposals that provide a more general signal that an entire web page is AI generated but nothing that labels only a section of a web page in a page that is otherwise authored by a human.

Weekly’s proposal acknowledges the reality that many web pages are partially AI generated. One example is the AI generated summaries of news content. The proposal specifically mentions news sites that contain a sidebar with AI generated summaries.

The proposal suggests creating an HTML attribute that can be applied at the section level using the

Weekly explains how it solves a problem:

“A news article page might contain a human-written investigation alongside an AI-generated summary sidebar. Existing approaches only support page-level disclosure (the tag proposed in whatwg/html#9479) or HTTP response-level signals (IETF draft-abaris-aicdh-00). Neither allows marking individual sections of a page, which is what 42+ commenters on the WHATWG issue identified as the key missing capability.

The EU AI Act Article 50 (effective August 2026) requires machine-readable marking of AI-generated text content, creating regulatory demand for exactly this kind of standard.”

The Aside Element Controversy

The

The definition of the

is:

“The

So the use of the

So, is this an imperfect solution in the context of an AI generated summary that is directly related to the document’s main content? I think it may be. Nobody in the GitHub discussion brings up this obvious disconnect in the use of the

The core rule of the

The Section Element Controversy

Using the

semantic HTML element is problematic because the role of this element is to group thematically related content together, with the key word being “thematic,” which refers to the subject matter, the theme, or topic of the content. It tells the browser or an assistive devices like a screen reader that everything within the section belongs to a specific topic.

The HTML spec for the

element is:

“A thematic grouping of content, typically with a heading.”

What a screen reader navigating a page that uses the

element sees:

• Introduction
• Methodology
• Results
• Conclusions
• Future Research

Using the the

element to declare the authorship or origin of the content breaks the intended purpose of the

element and could be problematic for people visiting a site with assistive devices.

Can Be Used On Any Element

The proposal said that the attributes can be used on any HTML element but the proposal seems to lean hard on the two semantic HTML elements discussed above.  Although none of the commenters in the proposal mentioned how the proposed use of the semantic HTML elements break the Accessibility Tree, this is an aspect worth discussing. Maybe it would have been better if the author had chosen different elements as examples but the proposal as it is right now leans  heavily on those two elements to do the heavy lifting.

Not A Settled Proposal

There is a lively conversation going on in the GitHub repository for the proposal. One of the purposes of the

They wrote:

“I’ve reviewed the proposal and the surrounding discussion, including the arguments in favor and against. However, the more I read, the more uncertainty I have about the practical necessity of introducing additional markup at the platform level. At the moment, this approach seems primarily aimed at satisfying formal or regulatory requirements, without a clearly demonstrated benefit for the web ecosystem as a whole.”

The takeaway is that the commenter sees the proposal as compliance-driven markup that platforms would be expected to add even when it does not clearly improve the web itself, and that concern becomes sharper if the implementation pushes disclosure into existing semantic HTML elements like