Artificial Intelligence and Assisted Work Policy
Verse & Variable
A line of verse. A line of inquiry.
لفظ سے لہر تک — Lafz se Lehar tak
1. Purpose
This policy establishes the principles governing the use of artificial intelligence (AI), machine learning systems, automated research tools, and other computational assistance in the creation and maintenance of Verse & Variable.
The publication explores literature, poetry, language, philosophy, cultural traditions, mathematics, mechanics, physics, and other scientific disciplines.
Developing such an interdisciplinary archive may involve tools capable of assisting with:
- Literary research.
- Linguistic analysis.
- Translation and transliteration.
- Historical and cultural research.
- Scientific exploration.
- Mathematical reasoning.
- Computational modelling.
- Writing and editing.
- Software development.
- Image and illustration generation.
- Information organisation.
- Accessibility improvements.
These tools can support the research and editorial process, but their outputs may contain errors, omissions, fabricated references, misleading interpretations, or unsupported conclusions.
This policy establishes the conditions under which such assistance may be used responsibly.
2. Fundamental Principles
2.1 Human Intellectual Responsibility
The person responsible for preparing and publishing an article remains accountable for its content.
An AI system cannot assume editorial responsibility, verify its own claims independently, or replace human judgement concerning the accuracy and appropriateness of published material.
2.2 Originality
AI assistance should not be used to disguise copying, reproduce another person’s protected work without appropriate authorisation, or misrepresent the origin of intellectual contributions.
Original commentary should remain genuinely independent and distinguishable from generated material and third-party sources.
2.3 Verification
Factual, historical, linguistic, scientific, mathematical, and bibliographic claims generated or suggested by AI must be checked against suitable sources before publication.
A confident answer is not evidence of accuracy.
2.4 Transparency
Material AI assistance should be disclosed where it meaningfully contributes to published content or affects how readers should understand its origin.
The disclosure should describe the actual assistance rather than rely on vague or misleading statements.
2.5 Intellectual Independence
AI tools may help explore ideas, but they should not determine the publication’s editorial direction.
Literary selection, interpretation, scientific judgement, and final publication decisions remain editorial responsibilities.
2.6 Respect for Original Creators
AI-assisted work must respect the authorship, copyright, cultural context, and intellectual contributions of writers, poets, lyricists, composers, translators, researchers, artists, and other creators.
2.7 Appropriate Use
The use of AI should serve a meaningful research, editorial, technical, or accessibility purpose.
Unnecessary automation should not be introduced merely because a tool makes it possible.
3. Scope of Application
This policy applies to AI and automated assistance used in:
- Selection and discovery of literary works.
- Original-text research.
- Author and contributor identification.
- Linguistic investigation.
- Translation and transliteration.
- Literary interpretation.
- Philosophical research.
- Historical and cultural research.
- Scientific explanations.
- Mathematical analysis.
- Research and citation discovery.
- Computational and numerical methods.
- Writing and editing.
- Website design and programming.
- Data organisation.
- Image generation.
- Scientific illustration.
- Search and archive development.
- Accessibility and technical maintenance.
It applies whether the assistance is provided by a standalone AI system, an integrated software feature, a machine-learning application, or another automated service.
4. Categories of AI Assistance
For editorial clarity, AI assistance should be distinguished according to its nature and significance.
4.1 Discovery Assistance
AI may help identify possible:
- Literary works.
- Authors and creators.
- Historical references.
- Research questions.
- Scientific concepts.
- Relevant academic publications.
- Alternative terminology.
- Further areas of investigation.
Such suggestions are starting points for research.
The existence, authenticity, and relevance of each important source must be verified independently.
4.2 Analytical Assistance
AI may assist with preliminary analysis of:
- Literary language.
- Metaphors and imagery.
- Linguistic structures.
- Historical context.
- Philosophical ideas.
- Scientific terminology.
- Mathematical formulations.
- Research methods.
The resulting analysis should be evaluated critically.
AI-generated interpretation should not automatically be presented as authoritative literary scholarship or established scientific knowledge.
4.3 Language Assistance
AI may support:
- Grammar checking.
- Spelling correction.
- Sentence refinement.
- Translation comparison.
- Transliteration exploration.
- Readability improvement.
- Terminology consistency.
Where language assistance substantially determines the wording or meaning of published material, appropriate review and disclosure are required.
4.4 Creative Assistance
AI may be used to develop preliminary:
- Article outlines.
- Editorial structures.
- Illustrative concepts.
- Visual compositions.
- Alternative phrasings.
- Explanatory examples.
Generated material should be reviewed before publication and should not be presented as an original literary work by a human creator unless its actual provenance supports that description.
4.5 Technical Assistance
AI may assist in preparing or reviewing:
- HTML.
- CSS.
- JavaScript.
- Liquid templates.
- YAML configuration.
- JSON and JSON-LD.
- Search-index generation.
- Data-processing scripts.
- Technical documentation.
- Accessibility improvements.
Code must be checked for correctness, security, licensing, compatibility, and unintended behaviour before deployment.
5. AI-Assisted Literary Research
Literary research requires particular sensitivity because language, historical context, and creative attribution cannot reliably be reconstructed from superficially similar online references.
5.1 Literary Discovery
AI may suggest:
- Poems.
- Ghazals.
- Nazms.
- Couplets.
- Songs.
- Literary traditions.
- Authors.
- Themes.
- Historical periods.
However, a suggested literary expression must be located in a reliable source before being considered for publication.
5.2 Verification of Original Text
AI-generated quotations should never be treated as authoritative original texts.
The selected expression should be checked against an appropriate source, such as:
- An authoritative published collection.
- An original recording or performance.
- A reliable literary archive.
- A verified manuscript or edition.
- A credible publisher or institutional repository.
This requirement is particularly important for short quotations that circulate widely with incorrect authorship or altered wording.
5.3 Literary Interpretation
AI may suggest possible readings of a literary expression.
These readings should be examined against:
- Original vocabulary.
- Grammar and syntax.
- Literary form.
- Cultural context.
- Historical usage.
- Relevant scholarly interpretation.
A generated interpretation should not be presented as the author’s actual intention without supporting evidence.
5.4 Cultural Context
AI-generated explanations of Indian and South Asian literary traditions may oversimplify complex cultural, regional, historical, philosophical, or devotional distinctions.
Verify important claims using appropriate language-specific, historical, and scholarly references.
Do not treat one cultural tradition as representative of every linguistic or regional community.
5.5 Original Literary Commentary
The four principal paragraphs of each individual exploration must remain original and intellectually responsible.
AI may assist with research or editorial refinement, but the final commentary should demonstrate genuine engagement with the selected literary work.
Generic generated observations should not be used merely to fill the required four-paragraph structure.
6. AI-Assisted Translation and Transliteration
Translation is one of the most demanding areas of AI-assisted literary work.
Machine translation may produce fluent text while incorrectly interpreting idioms, ambiguity, historical vocabulary, grammatical relationships, or poetic expression.
6.1 Preliminary Translation
AI may be used to generate a preliminary translation for comparison and linguistic investigation.
Such output must not automatically be treated as a verified translation.
6.2 Original-Language Verification
Where possible, compare the proposed translation with:
- The original script.
- Reliable dictionaries.
- Grammatical references.
- Historical language resources.
- Published translations.
- Scholarly commentary.
- Appropriate language-specific sources.
6.3 Ambiguity and Multiple Meanings
AI may select one meaning even when the original expression permits several interpretations.
Where ambiguity is meaningful, preserve it or explain the interpretative choice.
Do not allow a generated translation to introduce false precision.
6.4 Poetic Structure
Machine translation may lose:
- Rhythm.
- Meter.
- Rhyme.
- Metaphor.
- Repetition.
- Wordplay.
- Cultural associations.
- Emotional register.
Review these dimensions separately where they contribute materially to the original work.
6.5 Transliteration
AI-generated transliteration should be checked for consistency and linguistic plausibility.
Different transliteration systems may represent the same expression differently.
Do not describe one generated Roman-script representation as universally correct.
6.6 Published Translations
AI should not be used to disguise an existing copyrighted translation by changing selected words or sentence structures.
If an existing published translation informs the article, identify its source and comply with applicable rights and attribution requirements.
6.7 Disclosure
Where AI makes a material contribution to the published translation, disclose its role accurately.
For detailed requirements, also consult TRANSLATION_GUIDELINES.md.
7. AI-Assisted Historical, Cultural, and Philosophical Research
AI-generated historical and philosophical explanations require careful verification.
7.1 Historical Claims
Verify:
- Dates.
- Biographical information.
- Historical events.
- Literary movements.
- Publication details.
- Cultural practices.
- Geographical information.
- Institutional relationships.
Prefer original documents, reputable scholarly publications, established historical references, and authoritative archives.
7.2 Philosophical Attribution
Do not attribute an idea to a philosopher merely because an AI system associates it with similar terminology.
Where a philosophical concept is discussed:
- Identify the relevant thinker or tradition.
- Locate the original text where appropriate.
- Consult reputable scholarly interpretation.
- Distinguish historical meaning from contemporary usage.
- Acknowledge competing interpretations.
7.3 Indian Philosophical Traditions
Particular care is required when discussing concepts such as:
- Dharma.
- Karma.
- Moksha.
- Samsara.
- Rasa.
- Dhvani.
- Advaita.
- Bhakti.
- Vairagya.
These terms have context-dependent meanings across different texts, schools, and traditions.
Do not rely solely on an AI-generated explanation to establish their meaning.
7.4 Cultural Generalisations
Avoid using generated text that treats diverse Indian and South Asian linguistic traditions as homogeneous.
Regional, historical, linguistic, and philosophical differences should remain visible.
8. AI-Assisted Scientific Research
Scientific accuracy requires evidence appropriate to the nature of each claim.
AI systems may be useful for exploring technical subjects, but they can generate plausible explanations that contain subtle mathematical or physical errors.
8.1 Scientific Discovery
AI may help identify:
- Relevant disciplines.
- Technical terminology.
- Research questions.
- Possible physical mechanisms.
- Established theories.
- Related publications.
- Potential modelling approaches.
Important findings must be verified using appropriate scientific sources.
8.2 Scientific Definitions
Check generated definitions against authoritative textbooks, review papers, standards, and original research where appropriate.
Pay particular attention to terms that have different meanings across disciplines.
8.3 Equations and Mathematical Expressions
AI-generated equations must be checked independently.
Verify:
- Mathematical correctness.
- Variable definitions.
- Units.
- Dimensional consistency.
- Sign conventions.
- Boundary conditions.
- Initial conditions.
- Mathematical assumptions.
- Physical applicability.
Do not assume that a mathematically well-formatted equation is correct.
8.4 Nonlinear Mechanics and Wave Physics
Where AI assistance is used for subjects such as nonlinear vibrations, nonlinear elastic waves, harmonic generation, wave scattering, hysteresis, or contact nonlinearity, verify the actual governing principles.
Distinguish carefully between:
- Linear and nonlinear behaviour.
- Amplitude-dependent and amplitude-independent effects.
- Harmonic generation and ordinary linear propagation.
- Theoretical predictions and numerical observations.
- Physical mechanisms and descriptive analogies.
8.5 Metamaterials and Time-Dependent Systems
AI-generated explanations of mechanical metamaterials, phononic crystals, nonlinear cloaking, inverse design, and time-dependent systems should be checked against relevant research.
Do not assume that:
- Every periodic structure is a metamaterial.
- Every modulated system is a time crystal.
- Every reduction in scattering constitutes complete cloaking.
- Every inverse-design result is experimentally achievable.
- A numerical model demonstrates unrestricted physical performance.
The terminology must correspond to the actual system and its demonstrated behaviour.
8.6 Computational Modelling
AI may help explain numerical methods or prepare preliminary computational code.
However, model assumptions and numerical results require appropriate technical evaluation.
Where a simulation is discussed, examine:
- Governing equations.
- Constitutive relationships.
- Geometry.
- Boundary conditions.
- Mesh and discretisation.
- Numerical method.
- Solver configuration.
- Convergence.
- Sensitivity.
- Validation.
A generated simulation script does not establish that the physical model is correct.
8.7 Experimental Claims
Do not allow AI-generated explanations to turn numerical predictions into experimental findings.
Experimental claims must be supported by identifiable experimental evidence.
8.8 Scientific Uncertainty
Where the evidence does not establish a definitive conclusion, communicate that limitation.
AI-generated confidence must never determine the scientific status of a claim.
9. AI-Assisted Research References
Fabricated bibliographic information is a significant risk in AI-assisted research.
An AI system may produce convincing-looking references that do not exist, contain incorrect titles, or point to publications that do not support the associated claim.
9.1 Reference Verification
Every important reference suggested by an AI system must be checked.
Verify:
- Author names.
- Publication title.
- Journal or publisher.
- Publication year.
- Volume and issue.
- Page numbers.
- DOI.
- URL.
- Publication status.
9.2 DOI Verification
A DOI should be checked against its registered publication record.
Do not assume that a syntactically plausible DOI identifies a real article.
9.3 Relevance Verification
The existence of a publication does not establish that it supports a particular statement.
Consult the actual source and verify the relevant passage, figure, equation, or conclusion.
9.4 Peer-Review Status
AI-generated claims about a paper’s publication or peer-review status must be independently checked.
Distinguish between peer-reviewed publications, preprints, technical reports, theses, and conference presentations.
9.5 Bibliography Integrity
Do not include AI-generated references that have not actually been verified and consulted.
Refer to SOURCE_REGISTER.md and RESEARCH_STANDARDS.md.
10. AI-Assisted Writing and Editing
AI may help improve the readability and organisation of written material.
10.1 Permissible Assistance
Examples include:
- Grammar checking.
- Spelling corrections.
- Sentence restructuring.
- Reducing unnecessary repetition.
- Improving paragraph transitions.
- Checking terminology consistency.
- Preparing alternative formulations.
- Reviewing article structure.
10.2 Preservation of Editorial Voice
The publication’s writing should retain its established editorial character:
- Reflective.
- Mature.
- Natural.
- Intellectually curious.
- Culturally attentive.
- Precise.
- Substantive.
- Free from promotional exaggeration.
AI-generated language should not introduce generic or repetitive expressions that weaken the article.
10.3 Intellectual Contribution
AI assistance should not be used to create an appearance of original scholarly research where no such contribution has occurred.
The final article should distinguish independent analysis from sourced material and generated suggestions.
10.4 Unacceptable Practices
Do not:
- Present copied AI output as independently verified research.
- Fabricate quotations.
- Invent authorial intentions.
- Create false research findings.
- Invent references.
- Generate unsupported scholarly consensus.
- Conceal substantial unattributed borrowing.
- Use generated text to satisfy article length without meaningful intellectual content.
11. AI-Assisted Software Development
AI may assist with the development and maintenance of the publication’s technical infrastructure.
Potential areas include:
- Jekyll configuration.
- Liquid templates.
- HTML.
- CSS.
- JavaScript.
- YAML.
- JSON.
- JSON-LD.
- XML.
- RSS.
- Sitemap generation.
- Search indexing.
- Accessibility improvements.
- Technical documentation.
11.1 Code Review
Generated code must be examined before being committed.
Check:
- Syntax.
- Functionality.
- Compatibility.
- Accessibility.
- Performance.
- Error handling.
- Security.
- Unintended side effects.
11.2 Security
Do not deploy generated code that introduces:
- Unsafe HTML insertion.
- Unnecessary external scripts.
- Unvalidated input handling.
- Exposed credentials.
- Unintended data collection.
- Insecure dependencies.
- Unauthorised network behaviour.
Refer to SECURITY.md.
11.3 Dependencies and Licences
Check the licence and provenance of external code and libraries introduced through AI suggestions.
Do not assume that generated code is automatically free from third-party rights or licensing obligations.
Record significant external dependencies in ASSET_CREDITS.md.
11.4 Testing
A successful code suggestion is not evidence that the website works correctly.
Test the actual generated site after implementation.
Where applicable, verify:
- Jekyll builds.
- Page rendering.
- Base URL handling.
- Navigation.
- Responsive behaviour.
- Browser compatibility.
- Search functionality.
- Accessibility.
- Structured metadata.
11.5 Documentation
Where AI assistance results in a significant change to the publication’s architecture or functionality, update the relevant technical documentation.
12. AI-Generated Images and Visual Material
AI image-generation tools may be used for original visual resources where appropriate.
Such resources may include:
- Publication artwork.
- Conceptual illustrations.
- Decorative visual elements.
- Literary visual interpretations.
- Scientific diagrams.
- Explanatory illustrations.
12.1 Originality and Attribution
Generated images should be identified according to their actual origin where disclosure is materially relevant.
Do not attribute generated artwork to a historical artist, photographer, illustrator, or cultural institution unless that attribution is accurate.
12.2 Scientific Illustrations
Scientific visualisations require particular care.
An AI-generated conceptual illustration must not be presented as:
- An experimental photograph.
- A measured waveform.
- A laboratory observation.
- A validated simulation.
- An actual material sample.
- A research figure from an original publication.
Where the illustration is conceptual, identify it as such.
12.3 Historical and Cultural Material
Do not present generated depictions as authentic historical documents, original manuscripts, archival photographs, or verified representations of historical events.
If a visual is an artistic interpretation, make that distinction clear.
12.4 Text Inside Generated Images
Generated text within images may contain spelling errors, fabricated scripts, or incorrectly rendered characters.
For multilingual literary artwork, verify all visible script and wording.
Do not use an image as an authoritative reproduction of an original literary text when selectable and verified text can be provided separately.
12.5 Rights and Asset Records
Generated assets should still be evaluated for relevant platform terms, intellectual property concerns, and other applicable restrictions.
Record significant assets and their production context in ASSET_CREDITS.md where appropriate.
13. Confidentiality and Privacy
AI assistance must not compromise the confidentiality of unpublished, personal, restricted, or sensitive information.
Before submitting material to an external AI service, consider whether the material is appropriate to share with that service.
13.1 Restricted Material
Do not submit confidential material without appropriate authority.
This includes:
- Unpublished manuscripts belonging to another person.
- Private correspondence.
- Confidential institutional documents.
- Restricted research data.
- Unpublished scientific results belonging to others.
- Personal identification information.
- Sensitive personal information.
- Passwords and access credentials.
- Legally restricted material.
13.2 Personal Information
Avoid entering unnecessary personal information into external AI systems.
Where research material contains identifiable individuals, assess whether the information should be removed, anonymised, or handled through an appropriately protected process.
13.3 Service Policies
External AI providers may have different:
- Data-handling practices.
- Retention arrangements.
- Privacy settings.
- Security controls.
- Usage terms.
- Data-processing conditions.
Do not assume that every AI service provides identical privacy or confidentiality protections.
13.4 Human Responsibility
The person using the tool remains responsible for assessing whether the material can appropriately be shared.
Refer to PRIVACY.md.
14. Intellectual Property and Copyright
The use of AI does not remove the need to respect copyright, licensing, authorship, and other applicable rights.
14.1 Protected Literary Material
Do not use AI to reproduce or transform substantial protected poems, lyrics, translations, books, or other creative works without considering the applicable rights.
14.2 Unattributed Source Material
AI-generated paraphrasing should not be used to disguise the source of another person’s protected expression or ideas.
14.3 Generated Outputs
Do not assume that all generated outputs are automatically free of copyright concerns or unrestricted for every intended use.
Evaluate the relevant service terms and applicable rights.
14.4 Repository Licence
The repository’s MIT License applies according to its terms to material within its scope.
It does not automatically license independently protected literary works, images, recordings, translations, or other third-party assets.
14.5 Separate Asset Records
Where AI-generated or AI-assisted material is incorporated into the publication, identify its provenance and relevant usage conditions as appropriate.
Consult COPYRIGHT.md and ASSET_CREDITS.md.
15. Disclosure Requirements
The level of disclosure should reflect the significance of the assistance.
15.1 Assistance That May Not Require Individual Disclosure
Routine or incidental assistance may not require a separate public statement in every article.
Examples may include:
- Basic spelling checks.
- Routine formatting.
- Simple layout suggestions.
- Minor code-completion assistance.
- Non-substantive grammar corrections.
This does not exempt any factual or technical output from appropriate verification.
15.2 Assistance Requiring Consideration of Disclosure
Disclosure should be considered when AI materially contributes to:
- Original commentary.
- Substantial paragraph drafting.
- Literary interpretation.
- Translation.
- Scientific explanation.
- Mathematical analysis.
- Research synthesis.
- Generated illustrations.
- Original computational code.
- Significant editorial decisions.
15.3 Article-Level Disclosure
When AI assistance materially influences a particular article, the relevant disclosure should appear on that article or in an accessible accompanying disclosure section.
15.4 Publication-Level Disclosure
General information about the publication’s AI practices may be maintained in this policy.
A general policy should not replace article-specific disclosure when the actual contribution is materially different from ordinary routine assistance.
15.5 Accurate Description
Disclosures should explain:
- The type of assistance.
- The part of the work affected.
- Whether generated output was used directly or substantially revised.
- What verification was undertaken, where relevant.
Do not claim that every generated statement was independently verified unless that is actually true.
16. Recommended Disclosure Statements
The following examples provide suitable wording where they accurately describe the circumstances.
16.1 Preliminary Research
AI assistance: An AI-assisted research tool was used to identify preliminary research directions. Sources and factual claims included in the final article were independently checked against relevant references.
16.2 Language Editing
AI assistance: An AI-based language tool assisted with preliminary grammar and sentence refinement. The final article was reviewed and approved by its human author.
16.3 Translation Assistance
Translation disclosure: AI-assisted translation was used during the preliminary drafting process. The resulting text was checked against the original language and relevant linguistic references before publication.
16.4 Scientific Research Assistance
Research disclosure: AI tools assisted with preliminary exploration of scientific terminology and related concepts. The technical explanations and references were reviewed against identifiable scientific sources.
16.5 Programming Assistance
Technical disclosure: AI-assisted coding tools were used during development. The resulting implementation was reviewed and tested before being incorporated into the publication.
16.6 Generated Illustration
Visual disclosure: This conceptual illustration was generated using an AI-assisted image-generation process. It does not represent an experimental photograph or a validated scientific measurement.
16.7 Substantial Generated Drafting
AI assistance: An AI system contributed to the preliminary drafting of this section. The material was subsequently reviewed, revised, and verified by the responsible human author.
Use only the statement that accurately reflects the actual production process.
17. Unacceptable Uses
The following practices are inconsistent with the intellectual and editorial principles of Verse & Variable.
17.1 Fabricated Research
Inventing:
- Scientific experiments.
- Experimental results.
- Publications.
- Citations.
- DOI identifiers.
- Historical documents.
- Archival references.
- Author quotations.
17.2 Misrepresented Authorship
Presenting generated literary material as the original work of a named historical or contemporary creator.
17.3 False Verification
Claiming that an article, translation, reference, or scientific result has been independently verified when no such verification took place.
17.4 Concealed Substantial Assistance
Hiding material AI contributions where disclosure is necessary for readers to understand the origin of the published material.
17.5 Misuse of Scientific Authority
Presenting generated explanations or simulations as experimentally demonstrated facts without supporting evidence.
17.6 Cultural Misrepresentation
Using generated material that introduces unsupported cultural, historical, or linguistic claims without appropriate verification.
17.7 Copyright Evasion
Using AI paraphrasing or transformation to disguise unattributed borrowing or bypass applicable rights.
17.8 Privacy Violations
Submitting confidential or sensitive material to external systems without appropriate authority or consideration of the relevant privacy requirements.
17.9 Unreviewed Deployment
Publishing generated code, structured metadata, mathematical formulations, or other technical output without appropriate inspection and testing.
18. Human Review and Editorial Responsibility
AI-assisted material should undergo human review appropriate to its significance.
The degree of review should depend on the risk and complexity of the content.
Literary review
Check:
- Original expression.
- Language accuracy.
- Literary form.
- Cultural interpretation.
- Attribution.
- Translation decisions.
Scientific review
Check:
- Definitions.
- Equations.
- Mathematical assumptions.
- Governing principles.
- Experimental evidence.
- Numerical results.
- Technical terminology.
- Limitations.
Editorial review
Check:
- Originality.
- Coherence.
- Tone.
- Disclosure.
- Source attribution.
- Copyright considerations.
- Compliance with the four-paragraph commentary structure.
Technical review
Check:
- Code correctness.
- Compatibility.
- Security.
- Accessibility.
- Performance.
- Generated website behaviour.
Final Responsibility
The final decision to publish or retain AI-assisted content remains with the publication’s responsible human editor.
The existence of AI assistance should not diminish the standard of research, editorial judgement, or accountability expected of published material.
19. Tool and Model References
Where a particular AI system materially contributes to published work, the system may be identified in the relevant disclosure.
Useful information may include:
- Provider.
- Product or tool name.
- Model or version, where known.
- Purpose of use.
- Relevant date or period.
- Nature of the contribution.
Do not invent model identifiers or claim the use of a system that was not actually involved.
A tool reference should be included only when it improves transparency or is required by the relevant circumstances.
20. Record Keeping
For substantial AI-assisted work, maintain an appropriate internal record of the assistance used.
Depending on relevance, this may identify:
- Article identifier.
- Tool or system.
- Type of assistance.
- Material affected.
- Important verification steps.
- Disclosure decision.
- Relevant changes before publication.
It is not necessary to retain every prompt or intermediate generated response.
Records should be proportionate, securely managed, and consistent with the publication’s privacy and confidentiality responsibilities.
Do not preserve unnecessary sensitive information merely to create a detailed AI usage record.
21. Policy for Scientific and Computational Material
Because scientific subjects are an important part of Verse & Variable, additional safeguards apply to computational and technical content.
21.1 Governing Equations
AI-generated equations must be checked against an appropriate formulation or authoritative source.
21.2 Numerical Results
A generated numerical result must not be represented as an actual simulation output unless the simulation was genuinely performed.
21.3 Experimental Claims
An AI-generated description must not create the impression that an experiment took place when it did not.
21.4 Code Execution
Where generated code is used for analysis, inspect the inputs, outputs, dependencies, and assumptions.
Do not claim computational validation merely because code runs without producing an error.
21.5 Reproducibility
Where a computational result is important to an article, provide sufficient methodological information for its context and limitations to be understood.
21.6 Scientific Illustrations
Distinguish generated conceptual diagrams from actual physical measurements, experimental photographs, and validated numerical visualisations.
22. AI Use and the Four-Paragraph Editorial Structure
Every individual literary exploration must contain exactly four substantial paragraphs of original English commentary.
AI assistance must not turn this requirement into a mechanical text-generation exercise.
The four paragraphs should demonstrate:
- A considered encounter with the literary work.
- Deeper literary, philosophical, or cultural interpretation.
- A meaningful interdisciplinary reflection where appropriate.
- A thoughtful synthesis and concluding perspective.
AI-generated text should not be accepted merely because it fills all four positions.
The final commentary should remain coherent, original, independently reviewed, and intellectually meaningful.
Scientific connections must remain optional when they are not justified by the literary material.
23. Policy Compliance Checklist
Before publishing material involving AI assistance, review the following.
Provenance
- The type of AI assistance is understood.
- Original and generated material are distinguishable.
- Third-party contributions are identified.
- Relevant tool information is accurate.
Accuracy
- Factual claims are checked.
- Quotations are verified.
- Historical and cultural information is supported.
- Translation has been reviewed against the original.
- Scientific and mathematical claims are verified.
- References have been checked independently.
Intellectual integrity
- Originality has been preserved.
- Authorship is represented accurately.
- No fabricated citations or results are included.
- Interpretations are not presented as established facts without support.
- Substantial AI assistance is disclosed where appropriate.
Rights and privacy
- Copyright concerns have been considered.
- Third-party materials are properly attributed.
- Tool terms have been considered where relevant.
- Confidential information has been protected.
- Personal information has not been unnecessarily exposed.
Technical and editorial quality
- Generated code has been reviewed.
- Technical output has been tested.
- Images and diagrams are accurately described.
- Accessibility has been considered.
- The article meets the publication’s editorial structure.
- Final human editorial approval has been completed.
24. Relationship with Other Publication Documentation
This policy should be read alongside:
DISCLOSURE_POLICY.mdEDITORIAL_POLICY.mdEDITORIAL_WORKFLOW.mdRESEARCH_STANDARDS.mdSOURCE_REGISTER.mdTRANSLATION_GUIDELINES.mdLANGUAGE_STYLE_GUIDE.mdGLOSSARY.mdCOPYRIGHT.mdASSET_CREDITS.mdPRIVACY.mdSECURITY.mdCORRECTIONS_POLICY.mdDISCLAIMER.md
This policy establishes the specific responsibilities associated with AI-assisted work, while the other documents govern the corresponding editorial, research, linguistic, legal, and technical requirements.
25. Continuing Commitment
Artificial intelligence may expand the possibilities available to an independent literary and intellectual publication.
It may assist in exploring languages, discovering research directions, understanding unfamiliar scientific ideas, developing technical infrastructure, and presenting information in new ways.
However, technological capability should never be confused with intellectual authority.
The value of Verse & Variable must continue to arise from careful literary engagement, accurate research, meaningful interdisciplinary thinking, respect for original creators, and responsible editorial judgement.
AI should remain an instrument of exploration and assistance, not a substitute for intellectual integrity.
Verse & Variable
A line of verse. A line of inquiry.
لفظ سے لہر تک — Lafz se Lehar tak