AI Content Disclosures
Declare meaningful generative-AI use, confirm human review, and show readers clear authorship context.
ZyVOP lets every author declare how generative AI contributed to an article. As a ZyVOP transparency policy, an explicit choice is required when a new article is published or scheduled; drafts can remain undeclared while work is in progress. Google recommends useful disclosures where readers would reasonably expect creation context, but does not require this declaration or treat it as a ranking field.
Disclosure levels
| Level | Use it when |
|---|---|
| Human-written | A human wrote the article without meaningful help from generative-AI tools. Routine spelling, grammar, search, and non-generative tooling do not normally change this choice. |
| AI-assisted | A human led the work and AI materially helped with drafting, code generation, translation, research synthesis, or major edits. |
| AI-generated | An AI system generated most of the article, even if a human later edited or approved it. |
You can also add up to 500 characters of reader-facing context, such as which parts AI helped produce and what you verified.
Human review is required
AI-assisted and AI-generated articles cannot be newly published or scheduled until the author confirms that they reviewed the complete article and accept responsibility for its accuracy, sources, and rights.
In the editor
Open Publishing Settings → General → AI disclosure. Choose a level, add optional context, and—when AI materially contributed—confirm human review. If publication validation fails, the editor opens this panel, scrolls to the disclosure controls, focuses the missing control, and briefly highlights the section.
Published articles show a compact disclosure beside the article metadata. Existing articles published before this feature remain live and unchanged, but their next public update requires the author to declare a level.
REST API and frontmatter
Use these fields in Markdown frontmatter sent to POST /api/v1/articles:
ai_disclosure: ai_assisted
ai_disclosure_note: "AI helped reorganize the first draft; the author verified every claim and code sample."
ai_review_confirmed: trueAccepted ai_disclosure values are human_written, ai_assisted, and ai_generated. Camel-case aliases are also supported. Missing or invalid public-post declarations return an HTTP 400 validation error; GraphQL mutations return the equivalent bad-request error. Draft requests remain valid without a declaration.
CLI
The ZyVOP CLI validates the same fields locally for both developer-token and password-session publishing, including dry runs, and prints an actionable error before making a network request. With zyvop publish --local --devto, the CLI maps the declaration directly to DEV's native ai_disclosure_level field while keeping the provider API key local.
MCP
The zyvop_create_post tool accepts the same values through ai_disclosure, ai_disclosure_note, and ai_review_confirmed. ai_disclosure is required for schedule and publish modes. Draft creation can omit it.
DEV Community synchronization
When ZyVOP imports a DEV article, it preserves DEV's disclosure level. When ZyVOP cross-posts or updates an article on DEV, it maps the declaration to DEV's ai_disclosure_level field.
Search and policy context
Google's guidance focuses on useful, original, people-first content and recommends giving readers context about automation when it is reasonable to expect that context. Google does not document an AI-disclosure property for Article structured data, so ZyVOP keeps this as visible editorial metadata rather than inventing an unsupported schema field.
See Google Search guidance about generative AI content and Google's supported Article structured data fields.