اين خيلي مهم است
در آينده مهم‌تر هم
هر روز بيش از پيش
كار از كار مي‌گذرد
اگر به موقع اجباري نشود
الزامي
اين‌كه برچسب‌دار شوند
همه متون توليدي
آن‌چه ناشران مي‌نشرند

استاندارد كاملي نيافتم
كم و بيش پراكنده
اين شد كه يكي تهيه كردم
به فارسي البته
با هومص ترجمه به انگليسي
اين متن فراهم
اگر مورد توافق
اگر پايه براي همكاري
در كنوانسيون‌هاي بين‌المللي
اين مي‌تواند نجات محتواي انساني باشد
در مقابل گسترش محتواي هومصي

AI Disclosure Tag Standard ـ ADT-1.1

1. Purpose and Scope

This document establishes the AI-Generated Content Disclosure Tag (ADT-1.1), a mandatory, machine-readable, and legally-binding metadata standard for disclosing the involvement of artificial intelligence systems in the creation of scholarly, scientific, educational, and research outputs.
The standard applies to:
Peer-reviewed journal articles, preprints, and conference proceedings, Theses, dissertations, and academic books, Institutional research reports, white papers, and policy briefs, Any digital scholarly work submitted to a publisher, repository, or funding body that adopts this standard Compliance with ADT-1.1 is a condition of publication, indexing, and eligibility for research integrity assessments.

2. Guiding Principles

1. Mandatory Transparency ـ Every use of AI in the conception, drafting, revision, or sourcing of content must be declared, regardless of extent.
2. Strict Accountability ـ Human authors bear full, exclusive, and non-delegable legal and academic responsibility for all content, including AI-generated portions.
3. Verifiable Traceability ـ Each declared AI involvement must be corroborated by a verifiable human action (see Code "V" below), enabling audit and retrospective validation.
4. Non-circumvention ـ Omission, misrepresentation, or falsification of any AI contribution constitutes scientific misconduct and may result in retraction, institutional sanction, or legal liability under applicable research integrity frameworks.

3. ADT-1.1 Tag Structure ـ Revised with Compact Format

The disclosure tag shall appear exactly in the following format: {AIGEN: X,Y,Z,V}
Where:
"AIGEN" identifies the standard.
Letters denote specific AI assistance categories (see Section 4).
Code "V" is mandatory in all tags (see Section 4.6).
Codes are separated by commas (",") with no spaces between codes or after commas.

3.1. Placement

The tag must appear in both of the following locations, in identical form:
1. Immediately below the title of the work, on the first page or title screen; and
2. Within a dedicated section titled "Disclosure of AI Contribution" in the metadata, acknowledgements, or supplementary materials.
If the two locations conflict, the title-page tag prevails for extraction; the metadata section shall contain the exact same string.

3.2. Machine-Readability

The tag is designed for automated extraction. It is case-sensitive and must not contain any spaces (including spaces, tabs, or line breaks) inside the braces.
Standard extraction regex: "\{AIGEN:[A-Z,]+\}"
Explanation: This expression matches any string that starts with "{AIGEN:", followed exclusively by uppercase letters and commas, and ends with "}". Parsers must reject any tag containing spaces or treat it as malformed.

4. ADT-1.1 Category Codes

S > Structural planning AI used for outline, chapter hierarchy, argument flow, taxonomy, or classification. Does not include human-designed frameworks merely typed with AI assistance.
C > Content generation AI produced original prose, explanations, examples, calculations, or narrative segments that are incorporated substantively into the final work.
E > Editing and language refinement AI used for grammar, style, paraphrasing, reformatting, or error correction without altering substantive meaning.
H > Human review and validation One or more human authors have critically examined, fact-checked, corrected, and approved every AI-contributed element. This code is required if S, C, or E is present.
R > Citation of AI-tagged external sources The work cites at least one source that itself bears an AI disclosure tag (ADT-1 or later). This applies only if that source"s AI contribution materially influenced the citing work"s content.
V > Verified human confirmation Mandatory. Indicates that a named human author has submitted a verifiable attestation (see Section 4.2) confirming the accuracy of the entire tag.

4.1. Order rule

Codes must be listed in alphabetical order for consistency. "V" shall always be last.
Correct example: "{AIGEN: C,E,H,R,S,V=ADT-V-4F3A}"
Incorrect example: "{AIGEN: S, C, E, H, V}" (spaces are prohibited)

4.2. Verification Mechanism ـ Code "V"

To address the falsification risk, Code "V" imposes a two-layer verification:
1. Attestation Document ـ Each author must sign (digitally or physically) a brief declaration stating:
"I have personally reviewed the AI contributions declared in the ADT-1.1 tag and confirm that the description is complete and accurate. I accept full responsibility for the content."
This document must be deposited with the publisher or institution.
2. Verification Token ـ The publisher or repository shall generate a unique, time-stamped, non-transferable confirmation ID (e.g., "ADT-V-[UUID]") linked to the authors" ORCIDs. This token is appended to the tag as "V=token", not merely "V".
Example final tag with compact format: "{AIGEN: C,E,H,R,S,V=ADT-V-4F3A-9D2E}"
This token can be verified against the publisher"s registry during audits. Works bearing Code "V" without a valid token are considered non-compliant and will be flagged.

5. Examples of Compliant Tags

AI used only for spell-check; human reviewed all changes; verification provided > "{AIGEN: E,H,V=...}"
AI helped with outline and drafting, thoroughly revised by authors, with verification > "{AIGEN: C,E,H,S,V=...}"
As above, plus citing an AI-generated paper from a third party > "{AIGEN: C,E,H,R,S,V=...}"
No AI involvement at all No tag required, but publishers may ask for "{AIGEN: H,V=...}" to confirm human-only authorship.

6. Ethical and Legal Responsibilities

1. Authors must ensure AI usage complies with all institutional, national, and international laws, including data privacy, intellectual property, and export control regulations.
2. AI systems shall not be listed as co-authors under any circumstances.
3. Authors are jointly and severally liable for any inaccuracies, plagiarism, or ethical breaches in the work, irrespective of AI involvement.
4. Falsification of the ADT-1.1 tag (including false "V" attestations) constitutes gross academic misconduct and shall be reported to the authors" institutions and funding agencies.
5. Publishers are responsible for verifying the presence of a valid "V" token before final acceptance. They must retain the attestation document for a minimum of 10 years.

7. Institutional Implementation

To ensure worldwide adoption and enforcement, the following levels are recommended:
Authors Attach ADT-1.1 tag; sign and deposit attestation; ensure "V" token is obtained.
Peer reviewers Check tag consistency with content; flag suspicious mismatches.
Publishers / Editors Make ADT-1.1 a mandatory submission requirement; reject manuscripts without a valid tag; maintain token registry.
Indexing databases (Scopus, Web of Science, Crossref) Include tag as a searchable metadata field; permit filtering and bibliometric analysis by AI-use category.
Funding bodies Request the tag in final reports; use it for policy evaluation and grant compliance checks.

8. Auditing and Dispute Resolution

Random Audits: Publishers shall conduct randomised post-publication audits (≥5% of accepted papers) to validate the "V" attestation against the submitted manuscript. Whistleblower Mechanism: A confidential channel shall exist for reporting suspected tag falsification.
Sanctions: Proven violations will result in: Immediate retraction (with public notice) 5-year ban on publication in the adopting publisher"s journals Notification to the author"s institution and relevant funding bodies.

9. Version Control and Future Revisions

ADT-1.1 is a stable baseline. Future versions (1.2, 2.0) may introduce: Domain-specific sub-codes (e.g., for medical imaging, computational simulations) Multi-lingual tag variants preserving machine-readability Automated cross-checking with AI-detection tools (e.g., text watermarking) Blockchain-based immutable verification ledgers for "V" tokens All amendments shall be proposed through an open, international consultative process and ratified by a governing board comprising representatives from COPE, ICMJE, major publishers, and research integrity bodies.

10. Effective Date and Transition

ADT-1.1 supersedes ADT-1.0 as of [1 January 2027].
Works submitted under ADT-1.0 prior to that date remain valid, but publishers are encouraged to request migration to 1.1 during revision cycles.
Full compliance with Code "V" becomes mandatory for all new submissions after [1 July 2027].

End of Standard