Show HN: Declare AI – Open Standard for AI Content Disclosure

A new open-source project is proposing a universal standard for disclosing AI-generated content. “Declare AI” aims to solve the growing problem of distinguishing human from machine-created content across the internet.

The initiative comes as AI-generated text, images, and video become increasingly indistinguishable from human creations. Without disclosure standards, misinformation and trust erosion could accelerate.

The Problem

AI content is everywhere—and often undisclosed:

| Content Type | AI Share | Disclosure Rate |
|————–|———-|—————–|
| Blog posts | 30-40% | <10% | | Social media | 20-30% | <5% | | News articles | 10-20% | <15% | | Academic papers | 5-15% | <20% | | Marketing copy | 50-70% | <10% |

The gap between AI usage and disclosure is widening.

The Solution

Declare AI proposes a simple metadata standard:

“`json
{
“declare_ai”: {
“generated”: true,
“model”: “claude-sonnet-4”,
“human_edit”: “substantial”,
“timestamp”: “2026-02-25T10:00:00Z”
}
}
“`

The standard includes:

• Generation flag: Was AI involved?
• Model info: Which AI system was used
• Human involvement: None, minimal, substantial, or primary
• Timestamp: When content was created

How It Works

The standard can be embedded in multiple ways:

HTML Meta Tags

“`html

“`

JSON-LD Schema

“`html

“`

API Headers

“`
X-Declare-AI: generated=true; model=claude; human=substantial
“`

Watermark Integration

For images and video, the standard integrates with existing watermarking systems like C2PA.

Industry Support

The project has early backing from:

| Organization | Role | Commitment |
|————–|——|————|
| Mozilla | Technical advisor | Browser integration |
| Creative Commons | Standards body | Endorsement |
| Automattic | Early adopter | WordPress plugin |
| Substack | Evaluating | Potential integration |
| Medium | Watching | Pending |

The goal is broad adoption before regulation mandates it.

Why Now?

Several factors are driving urgency:

Regulatory Pressure

• EU AI Act: Requires AI content labeling
• US state laws: California, New York considering disclosure requirements
• Platform policies: Google, Meta updating content rules

Trust Erosion

Surveys show declining trust in online content:

• 60% of users worry about AI misinformation
• 45% can’t distinguish AI from human content
• 70% support mandatory disclosure

Competitive Dynamics

Companies that disclose AI use may gain trust advantages. Early adopters could differentiate themselves.

Technical Challenges

Implementing the standard faces obstacles:

Detection Accuracy

Current AI detectors have high false positive rates. The standard relies on voluntary disclosure, not detection.

Enforcement

Without central authority, compliance is voluntary. Bad actors may ignore the standard entirely.

Complexity

Different content types (text, image, video, audio) need different disclosure mechanisms. Harmonizing these is technically challenging.

Legacy Content

Billions of pages exist without AI disclosure. Retrofitting is impractical at scale.

Key Takeaways

• Project: Declare AI proposes open standard for AI content disclosure
• Metadata: Generation flag, model info, human involvement, timestamp
• Integration: HTML meta tags, JSON-LD, API headers, watermark systems
• Support: Mozilla, Creative Commons, Automattic backing
• Drivers: EU AI Act, US state laws, platform policies, trust concerns
• Challenges: Detection accuracy, enforcement, complexity, legacy content
• Goal: Broad adoption before regulation mandates it

The Bottom Line

Declare AI represents a proactive approach to a problem that regulators will inevitably address. By establishing a voluntary standard now, the project hopes to shape eventual requirements rather than having them imposed.

The technical design is sound. The metadata is simple, the integration paths are clear, and the standard is flexible enough for different content types. But technical merit alone won’t drive adoption.

Success depends on industry coordination. If major platforms adopt the standard, it becomes de facto mandatory. If they don’t, it remains a niche specification for ethics-conscious publishers.

The timing is right. Regulatory pressure is building, trust is eroding, and companies need a solution. Declare AI offers one. Whether it succeeds depends on whether the industry values self-regulation over government mandates.

FAQ

What is Declare AI?

Declare AI is an open-source project proposing a universal standard for disclosing AI-generated content. It uses metadata to indicate whether content was AI-generated, which model was used, and the level of human involvement.

How does the standard work?

The standard embeds disclosure metadata in HTML meta tags, JSON-LD schema, API headers, or watermark systems. The metadata includes generation status, model information, human involvement level, and creation timestamp.

Who supports Declare AI?

Early supporters include Mozilla (technical advisor), Creative Commons (standards endorsement), and Automattic (WordPress plugin). The project aims for broad industry adoption before regulations mandate disclosure.

—

Sources: Hacker News Discussion, GitHub Repository, Creative Commons

Tags: Declare AI, AI Disclosure, Content Standards, Open Source, AI Ethics, Metadata