Use of AI in Research

The Arab Center for Research and Policy Studies (ACRPS) acknowledges that artificial intelligence tools have become an integral feature of the contemporary research and knowledge production environment. The ACRPS recognizes the opportunities these tools offer to accelerate certain tasks and to enhance the results. Conversely, the ACRPS also recognizes the ethical, legal, and scholarly risks arising from the irresponsible or unregulated use of AI – particularly regarding originality, academic integrity, the accuracy of information, intellectual property rights, confidentiality, and data protection.

This policy complements the ACRPS’s established Publishing Ethics and is to be read in conjunction with them, not in their stead. It outlines specific guidelines for the use of AI across the ACRPS’s various activities and publications. The ACRPS does not consider the use of AI in itself a breach of research ethics, nor as grounds to question the originality of research. Rather, the legitimacy of such use is determined by its nature and extent, as well as its adherence to academic and ethical standards.

Depending on the nature of the activity, this policy applies to research papers, studies, books and book chapters, essays, research proposals, abstracts, conference and seminar papers, policy papers, situation assessments, political and economic analyses, research unit publications, translations, editing and proofreading work, peer review reports, and other materials published or produced by the Center or submitted to it for consideration.

Originality and Quality: Authors, researchers, translators, and other contributors must ensure that work submitted to the ACRPS is the product of original, human intellectual contribution and is free from the plagiarism of ideas or text from any source – including content generated by AI tools – and adheres to professional and ethical research standards. Such tools must not replace the author’s substantive intellectual contribution, particularly regarding the construction of arguments, the formulation of questions and hypotheses, the interpretation of results, the drawing of conclusions, or the making of scholarly or analytical judgements. When evaluating any contribution, the ACRPS focuses on its scholarly merit, originality, quality, and epistemic value; the use or suspected use of AI tools does not constitute independent grounds for rejecting the work.
Human Responsibility: The author bears full responsibility for every part of the work, including any content, data, translations, or analyses produced with the assistance of AI tools. These tools may not be designated as authors, co-authors, translators, or reviewers; nor may scholarly or ethical responsibility be attributed to them. Authors must verify all outputs from such tools and conduct any necessary human review. Disclosing the use of such tools does not absolve the author of responsibility for the work’s accuracy and its scholarly and legal integrity.
Legal and Ethical Obligations and Intellectual Property Rights: Users of AI tools must respect intellectual property and copyright laws throughout all stages of preparing their work. This applies both to materials submitted to these tools and to the outputs generated by them. Where necessary, users must adhere to the tools’ terms of use and their associated privacy and intellectual property policies. They must ensure that the use of these tools does not conflict with the rights of the author, the ACRPS, or other holders of rights – including issues regarding the retention by service providers of input materials or their use in model development. The user bears responsibility for any legal or ethical violation resulting from the illicit use of these tools.
Confidentiality, Data Protection, and Unpublished Materials: Users must not enter confidential, sensitive, or unpublished materials into external AI tools unless expressly permitted by the ACRPS’s approved policies and in keeping with adequate data protection safeguards. This applies to manuscripts, research proposals, peer review reports, books, unpublished translations, internal documents, interviews, research data not in the public realm, personal information regarding research participants, and any materials subject to confidentiality agreements or contractual restrictions. For research involving personal data, interviews, or fieldwork, users must strictly adhere to the relevant ethical approvals and refrain from transferring data to external platforms in any manner that goes beyond the limits of such approvals.
Accuracy of Information, References, and Academic Transparency: Users must consider the risks associated with using AI tools, including the fabrication of references; the mixing of accurate and inaccurate data; the attribution of ideas, results, or quotations to sources where they do not appear; the generation of literature reviews that conflate studies or attribute inaccurate positions or findings to them; the paraphrasing or translation of texts to conceal plagiarism; or the presentation of pre-existing work as new. It is incumbent on the author to personally verify the existence of sources and the accuracy of the data being cited, as well as to confirm that sources actually contain the ideas and quotations attributed to them. Authors must not rely on AI tools for academic or factual information instead of consulting original or authoritative sources, except in cases where the tool or its output is itself the subject of the research or analysis.
Disclosure of AI Usage: Authors and other contributors must disclose any use of AI tools in the preparation of their work. Depending on the nature of the work, this may include text generation or paraphrasing; translation; the summarization or classification of literature; transcription of interviews; organization of references or citations using generative tools; data collection or analysis; programming; the preparation of tables, figures, and audiovisual materials; and any other use that substantially impacts the work’s content, methodology, or formulation. The disclosure must include, as appropriate, the name and version of the tool (if possible), the nature of the tasks performed with its assistance, the extent of the author’s reliance on it, and the nature of human oversight and verification of its outputs. Disclosure is not required for use limited to simple spelling or grammar correction, technical functions that do not generate new content or rephrase ideas, or the use of AI tools as search engines (as with traditional search engines) to find information and sources, provided such use does not stray into relying on content generated by that tool in order to prepare the research.
AI Usage in Research, Data, and Methodology: AI tools may be used to assist in organising research materials; transcribing, coding, classifying, and analysing data; programming; or performing calculations and quantitative analyses, whenever methodologically appropriate, provided this does not conflict with the regulations set out in Clause 4 regarding confidentiality and the protection of data and unpublished materials. Researchers must disclose such use and independently verify the validity of the tool’s output. If AI tools form part of the research methodology or could influence its results, researchers must explain their role in a manner that clarifies how the results were reached and allows for their evaluation while maintaining an appropriate record of how such tools were used. Researchers must not fabricate data, results, or statistics nor present artificial data or AI-generated simulations as genuine field or experimental data. Furthermore, valid data must not be interpreted in a manner inconsistent with their nature or the methodology employed; the researcher bears responsibility for ensuring the consistency of data, analyses, and tables with the results and conclusions presented in the work.
Research Proposals and Abstracts: Authors must not use AI tools to prepare research proposals and abstracts for submission to the ACRPS. This applies to proposals and abstracts submitted for conferences, seminars, calls for papers, special issues of journals, and other scholarly activities. The proposal or abstract must be entirely the author’s own work, reflecting their actual conceptualization of the topic, research question, methodology, research material, and familiarity with the relevant literature. Moreover, the proposal or abstract must not contain data, results, or conclusions that imply that research has been conducted or that data has been collected or analysed, when the work is in fact still in the conceptual stage or relevant phases of the research have yet to be completed.
Translation: AI tools may be used for translation purposes to assist in understanding texts, terminology, or sentence structures, or to compare alternatives and select the most appropriate translation – much like digital dictionaries and machine translation (MT) tools. Translators must not upload a source text to an AI tool, whether in full or in substantial extracts that would allow its reconstruction, unless expressly authorized by the publisher or rights holder. Translators must not rely on AI tools to produce the translation itself nor present a machine–generated translation as a finished product. The translator bears full responsibility for the accuracy, integrity, and style of the translated text and its fidelity to the original.
Editing and Proofreading: Editors may use AI tools for editing and proofreading in a limited and specific manner, such as seeking assistance with editing a particular sentence or phrase, seeking guidance on language, grammar, or style matters, or comparing specific phrasings. Editors must not upload entire files or manuscripts, or substantial portions thereof, to AI platforms for automated editing or proofreading. For documentation purposes, however, footnotes or reference data may be entered, separately, for formatting according to the relevant citation style.
Handling Academic Comments: Authors must not use AI tools to process evaluation, peer review, or scholarly editing comments. This includes uploading such comments to an AI tool or using it to summarize or interpret them, propose ways to address them, draft responses to them, or implement the requested revisions. Any academic response or addressing of such observations must stem from the author’s own understanding, judgement, and knowledge of the subject matter.
Images, Tables, Figures, and Audiovisual Materials: AI tools must not be used to generate tables, figures, images, maps, or recordings that are presented as being based on original data or sources, unless they can be traced back to – and verified against – those data or sources. Furthermore, visual and audiovisual materials must not be altered in a way that changes their meaning or misleadingly claims to document an event or original material. If AI tools are used substantially in the production or modification of such materials, this must be clearly disclosed. This prohibition does not apply to declared and justified methodological uses in which the tools themselves form part of the research subject or methodology, provided that such use, its nature and scope are clearly disclosed.
AI Usage in Peer Review, Evaluation, and Academic Decision–Making: Reviewers must not use generative AI tools to evaluate manuscripts, papers, or research proposals submitted to them, nor to draft reviewer reports. Reviewers must not upload submitted work or accompanying materials, in whole or in part, to such tools in order to safeguard confidentiality, intellectual property rights, and the independence of the review process. The review report must be the product of the reviewer’s own personal effort and academic expertise. The same principle applies, depending on the nature of the task, to academic committees, proposal and award evaluation committees, and other processes requiring independent academic judgement. Furthermore, academic or editorial decisions related to the acceptance, rejection, prioritization, or withdrawal of work must not be delegated to AI tools.
Suspicion of Undisclosed or Inappropriate Use: If evidence arises suggesting the use of AI tools in a manner that may violate this policy, the relevant party at the ACRPS reserves the right to request clarification from the author regarding the nature and extent of such use, as well as verification of references, citations, data, disclosures, and other relevant elements. Linguistic style, formatting, variations in drafting quality, or results from AI-generated text detection tools must not, on their own, be treated as sufficient evidence of a violation or as grounds for rejection; any action taken must be supported by verifiable evidence concerning the integrity or the scholarly integrity of the work.
Violations and Associated Measures: Any breach of any of these rules constitutes a violation of the ACRPS’s approved ethical and academic standards. This applies to instances of plagiarism or deception; fabrication of information, data, or references; breaches of confidentiality or intellectual property rights; materially incorrect or incomplete disclosures; and any other practices affecting the integrity or the scholarly integrity of the work – whether through AI tools or not. The ACRPS is responsible for distinguishing between errors or shortcomings that can be corrected, and violations that warrant institutional action. Depending on the nature, severity, and stage of discovery of a given case, the ACRPS reserves the right to request clarification, correction, or revision; to halt the processes of peer review or publication; to reject, correct, or retract a contribution (even after publication); or to take any other measures it deems appropriate in accordance with its approved regulations.

Given the rapid evolution of AI tools and their usage methods, this policy is subject to periodic review and revision to align with relevant technological developments as well as applicable academic, ethical, and legal standards. ACPRS publications, periodicals, units, programmes, and conferences may adopt additional, more specific guidelines if the nature of their activities so requires, provided that such guidelines remain consistent with the general principles set forth in this policy.