AI Policy

Artificial Intelligence (AI) & Generative AI Policy

Policy Statement

MANUSKRIPTA, published by Masyarakat Pernaskahan Nusantara (Manassa) is committed to the ethical, transparent, and responsible use of artificial intelligence (AI) in philological research and scholarly publishing.

AI-assisted technologies may support manuscript preparation or technical analysis. However, their use must not compromise originality, independent intellectual contribution, methodological transparency, intellectual property, confidentiality of private or community-owned manuscripts, or the reliability of research findings.

This policy applies to authors, peer reviewers, editors, Editorial Board members, and all other parties involved in the publication process. It aligns with the journal’s Publication Ethics and is informed by relevant guidelines from the Committee on Publication Ethics (COPE).

Core Principle:

Artificial intelligence may assist scholarly work, but it must not replace human intellectual contribution, critical philological analysis, textual interpretation, or academic responsibility. Human authors, reviewers, and editors remain fully accountable for the content and decisions within their respective roles.

Key Summary

  • Permitted: Limited supporting uses (language editing, formatting assistance, technical script analysis coding).
  • Required: Honest and transparent disclosure of AI usage.
  • Prohibited: Fabrication of manuscript texts, creation of artificial transliterations/sources, and violation of manuscript ownership confidentiality.
  • Accountability: Full scientific responsibility rests entirely on human authors.
  1. Permitted Use of AI by Authors

Authors may use AI-assisted technologies for limited supporting activities that improve the presentation, linguistic clarity, or technical preparation of manuscripts without replacing their own philological analysis.

Permitted uses include:

  • Language Editing: Correcting grammar, spelling, punctuation, and linguistic terminology consistency.
  • Readability: Enhancing sentence structure without altering the meaning of philological or codicological findings.
  • Outline Assistance: Assisting in organizing sections or structuring manuscript drafts.
  • Formatting Support: Assisting with technical consistency in citations and references.
  • Translation Assistance: Supporting translation, provided that technical terminology (e.g., codicological and philological terms) is independently verified by the authors.

Note: All AI-assisted material must be critically reviewed, verified, and validated by human authors. AI suggestions should never be accepted without verifying textual evidence, transliterations, and manuscript citations.

  1. AI as a Research Method or Subject

MANUSKRIPTA recognizes that AI may itself serve as a research subject or a documented methodological tool in digital humanities, digital philology, optical/handwritten text recognition (OCR/HTR) for ancient manuscripts, or stylometric analysis of Nusantara texts.

Where AI forms part of the research methodology, authors must explicitly report:

  1. The tool, model, or software used (including version and access date).
  2. The specific purpose and methodological rationale for using AI in the manuscript analysis.
  3. The input data, manuscript corpora, or digital images processed.
  4. How outputs were generated, selected, validated, and philologically analyzed.
  5. The extent of human oversight and intervention.
  6. Methodological limitations, potential AI reading biases, and output accuracy evaluations.
  1. Prohibited Use of AI

The journal strictly prohibits AI practices that compromise the authenticity, originality, or scientific integrity of manuscript research.

Unacceptable practices include:

  • Presenting substantial AI-generated content as independent human intellectual work.
  • Using generative AI to replace authors' critical philological analysis, textual interpretation, reconstruction of textual genealogy (stemma codicum), or drawing conclusions.
  • Fabricating or falsifying manuscript readings, codicological data, copying history, or textual evidence.
  • Generating non-existent citations, manuscript catalogues, shelfmarks, or references.
  • Manipulating digital manuscript images in a way that misrepresents the physical reality of the artifact.
  • Using AI to conceal plagiarism or manipulate similarity index scores.
  • Uploading confidential or restricted private/community manuscript data or images into public AI tools without official authorization from manuscript owners.
  1. AI Disclosure and Transparency

Authors must accurately and transparently declare the use of generative AI tools, chatbots, large language models (LLMs), custom HTR/OCR models, or language assistants used during manuscript preparation.

The declaration must state:

  • The name of the tool and developer (e.g., ChatGPT-4o, Claude, Transkribus, etc.).
  • The specific purpose of its use.
  • Where and how it was employed in the article.
  • The manner in which authors verified and validated the resulting outputs.

The statement must appear in the cover letter upon submission and, where relevant, within the Methods or Acknowledgments section of the manuscript.



Example of an AI Use Statement:

"The authors used [AI Tool Name] for [specific purpose, e.g., English language editing / preliminary HTR processing] in [specific section(s)]. All AI-generated outputs were independently reviewed, cross-checked, and verified against the original physical/digital manuscript by the authors. The AI tool was not used to fabricate manuscript data, interpret texts, or draw philological conclusions. The authors retain full responsibility for the accuracy, originality, and integrity of this manuscript."

Example When AI Was Not Used:

"The authors declare that no generative AI or AI-assisted technologies were used in the preparation, analysis, or writing of this manuscript."

 

  1. AI, Authorship, and Human Responsibility

AI tools, chatbots, or large language models must not be listed as authors or co-authors on manuscripts submitted to MANUSKRIPTA.

Authorship requires substantive human intellectual contribution, final manuscript approval, and full accountability for research integrity. Human authors remain solely responsible for:

  • The originality and accuracy of manuscript content, readings, and transliterations.
  • The authenticity of codicological data and manuscript provenance.
  • Accurate citations, variant reading comparisons, and bibliographies.
  • Compliance with manuscript protection ethics and authorization from manuscript owners.
  1. Ethical Safeguards and Data Confidentiality

MANUSKRIPTA publishes scholarly works that frequently involve sacred texts, private collections, rare manuscripts, and indigenous cultural traditions.

The following materials must not be uploaded to public AI systems (open AI models) without written authorization and effective security safeguards:

  • Transcripts or digital images of unpublished or restricted private manuscript collections.
  • Sensitive personal details regarding manuscript owners or protected local family traditions.
  • Fieldwork data or interview records subject to confidentiality agreements (informed consent).
  1. Use of AI by Peer Reviewers

Peer reviewers must maintain strict confidentiality regarding manuscripts under evaluation.

Reviewers must not:

  • Upload unpublished manuscripts or excerpts into AI tools for analysis, summarization, or evaluation.
  • Use AI to generate entire peer review reports without independent, critical human evaluation.

Reviewers are fully responsible for the objectivity, accuracy, and philological soundness of the review reports they provide.

 

Commitment to Academic Integrity:

MANUSKRIPTA supports technological advancements in manuscript studies and the adoption of modern digital tools, provided their application strictly adheres to scientific ethics, academic honesty, methodological transparency, and respect for Nusantara's manuscript heritage.