Introduction / 引言

Silence recognizes the increasing use of artificial intelligence (AI) technologies in scholarly research, academic writing, and editorial communication. AI-assisted tools may support particular research and writing tasks, but their use must remain consistent with academic integrity, transparency, intellectual responsibility, and applicable ethical and legal requirements.

《Silence》認識到人工智慧(AI)技術日益廣泛地應用於學術研究、論文寫作及編輯交流。人工智慧輔助工具可以支持特定研究與寫作工作,但其使用必須符合學術誠信、透明度、智識責任,以及適用的倫理與法律要求。

This policy establishes the journal's expectations concerning the disclosure, responsible use, and human oversight of AI-assisted technologies in manuscripts submitted to Silence. It also addresses the use of AI within research methodologies and the responsibilities of editors and reviewers when working with confidential scholarly materials.

本政策說明《Silence》對投稿過程中人工智慧輔助技術之使用披露、負責任使用及人類監督的要求,同時涵蓋人工智慧作為研究方法的情況,以及編輯與審稿人在處理保密學術材料時所應承擔的責任。

The journal distinguishes between AI assistance in manuscript preparation and AI technologies used as substantive components of research design, data processing, analysis, or modeling. The information needed to evaluate these uses may differ, but authors remain responsible for accurately describing the role of the technology and the reliability of the resulting scholarly work.

本刊區分用於稿件準備的人工智慧輔助,以及作為研究設計、資料處理、分析或建模之實質組成部分的人工智慧技術。不同用途所需的說明可能有所差異,但作者均須準確描述技術所扮演的角色,以及最終學術成果的可靠程度。

AI-assisted technologies may include large language models, generative AI applications, automated writing systems, image-generation tools, AI-based analytical software, and other systems that generate, transform, classify, or interpret information through artificial intelligence techniques. The relevance of this policy depends on the function performed by the technology and its contribution to the research or manuscript, rather than on a particular commercial product name.

人工智慧輔助技術可能包括大型語言模型、生成式人工智慧應用、自動化寫作系統、圖像生成工具、人工智慧分析軟體,以及其他透過人工智慧技術生成、轉換、分類或解釋資訊的系統。本政策是否適用,取決於技術實際執行的功能及其對研究或稿件的貢獻,而非特定商業產品名稱。

The use of AI is not, by itself, evidence of academic misconduct. Equally, disclosure of AI use does not exempt a manuscript from the journal's requirements concerning originality, methodological rigor, factual accuracy, appropriate attribution, and publication ethics. The scholarly contribution must remain identifiable, appropriately supported, and subject to accountable human judgment.

使用人工智慧本身並不構成學術不端的證據。同樣,披露人工智慧的使用情況,也不會使稿件免於遵守本刊對原創性、方法論嚴謹性、事實準確性、適當來源標示及出版倫理的要求。學術貢獻仍須清晰可辨、具有適當依據,並接受可以承擔責任的人類判斷。

The distinction between technical assistance and substantive intellectual contribution is particularly important. Assistance with grammar or formatting may affect the presentation of a manuscript without determining its scholarly findings. By contrast, AI-generated interpretations, classifications, analyses, or research materials may directly influence the evidence and reasoning on which conclusions depend. The extent of disclosure, verification, and methodological explanation should reflect these differences.

技術性協助與實質智識貢獻之間的區別尤其重要。文法或排版協助可能影響稿件的呈現方式,卻不直接決定研究發現;相反,人工智慧生成的解釋、分類、分析或研究材料,可能直接影響結論所依據的證據與推理。披露、驗證及方法論說明的程度應反映這些差異。

This policy should be read together with the journal's Author Guidelines, Publication Ethics Policy, Plagiarism Policy, Research Data Sharing Policy, Reviewer Guidelines, Peer Review Policy, and Protection of Intellectual Contributions. The use of AI does not remove responsibilities established by those policies or create an exemption from the standards applicable to a particular article type.

本政策應與本刊的 Author Guidelines(投稿指南)、Publication Ethics Policy(出版倫理政策)、Plagiarism Policy(剽竊政策)、Research Data Sharing Policy(研究資料共享政策)、Reviewer Guidelines(審稿人指南)、Peer Review Policy(同行評審政策)及 Protection of Intellectual Contributions(智識貢獻保護)一併理解。使用人工智慧不會取消上述政策所確立的責任,亦不構成免於遵守特定文章類型學術標準的理由。

AI Tools and Authorship / 人工智慧工具與作者資格

Artificial intelligence systems, including large language models, generative AI applications, and automated writing tools, cannot be listed as authors or coauthors of manuscripts submitted to Silence. The same principle applies when AI technology has contributed substantially to drafting, generating images, processing information, or producing research outputs.

人工智慧系統,包括大型語言模型、生成式人工智慧應用及自動化寫作工具,不得被列為提交至《Silence》之稿件的作者或共同作者。即使人工智慧技術對文字起草、圖像生成、資訊處理或研究成果製作具有重大參與,亦適用相同原則。

Authorship entails intellectual accountability, responsibility for the accuracy and integrity of the work, approval of the submitted content, and the ability to respond to questions concerning the research. AI systems cannot independently assume these responsibilities, verify the legitimacy of the materials they produce, or fulfill the obligations associated with scholarly authorship.

學術作者資格意味著承擔智識責任、對作品的準確性與完整性負責、批准提交內容,以及能夠回應與研究有關的問題。人工智慧系統無法獨立承擔此類責任、核實其所產生材料的正當性,或履行學術作者資格所涉及的義務。

All listed authors must be human contributors who satisfy the journal's applicable authorship requirements. The use of AI-assisted tools does not replace the need for genuine intellectual contributions, appropriate attribution among human contributors, or collective responsibility for the submitted manuscript.

所有列名作者均須為符合本刊適用署名要求的人類貢獻者。使用人工智慧輔助工具不能取代實質智識貢獻、人類參與者之間的適當署名安排,或作者對所提交稿件共同承擔的責任。

Authors must not identify an AI system as a scholarly author in the manuscript byline, authorship metadata, or other formal authorship records. Describing an AI system as a research tool, software resource, or source of technical assistance is distinct from attributing authorship to that system. Human contributors remain responsible for decisions about research design, interpretation, manuscript approval, and the integrity of the final work.

作者不得在稿件署名欄、作者元資料或其他正式署名紀錄中,將人工智慧系統列為學術作者。將人工智慧系統描述為研究工具、軟體資源或技術協助來源,與賦予該系統作者資格屬於不同事項。人類貢獻者仍須對研究設計、研究解釋、稿件批准及最終作品的完整性負責。

AI Acknowledgment Is Not Authorship / 人工智慧使用說明不等於作者署名

Where an AI tool has assisted with manuscript preparation or research, its role should be described through the appropriate disclosure or methodological explanation. Identifying a tool and its contribution does not confer authorship status on the system, nor does acknowledgment transfer responsibility for its output away from the human authors.

如人工智慧工具協助稿件準備或研究工作,作者應透過適當的使用聲明或方法論說明交代其作用。標示工具及其貢獻,並不代表賦予該系統作者資格,也不會使相關輸出的責任由人類作者轉移至人工智慧系統。

Where multiple human contributors participate in AI-assisted work, their respective intellectual and practical contributions should be represented accurately under the journal's existing authorship and acknowledgment requirements. The use of AI should neither obscure substantive human contributions nor be used to justify honorary, guest, or otherwise inappropriate authorship.

當多名人類貢獻者參與人工智慧輔助研究時,應依據本刊既定的作者署名與致謝要求,準確呈現各自的智識及實務貢獻。人工智慧的使用既不應掩蓋人類參與者的實質貢獻,也不得被用來合理化榮譽作者、掛名作者或其他不適當署名安排。

Disclosure of AI Use / 人工智慧使用披露

Authors who use AI-assisted tools in preparing a manuscript must disclose that use transparently. The disclosure should describe the nature and purpose of the assistance and should allow editors and readers to distinguish the authors' scholarly contributions from activities supported by automated systems.

作者如在稿件準備過程中使用人工智慧輔助工具,必須透明披露相關使用情況。披露內容應說明輔助工作的性質與目的,使編輯及讀者能夠區分作者本身的學術貢獻與自動化系統所提供的協助。

Disclosure should be sufficiently specific to explain what the technology was used to accomplish. A general statement that AI was used, without describing the relevant task, may be inadequate when the assistance materially affected the manuscript's content, analysis, interpretation, or presentation. The disclosure should correspond to actual use rather than claim that a tool performed activities it did not undertake.

披露內容應具備足夠具體性,使讀者能夠理解技術實際完成的工作。當人工智慧協助實質影響稿件內容、分析、解釋或呈現時,僅籠統表示曾使用人工智慧,而未說明相關用途,可能不足以支持透明評估。披露應與實際使用情況相符,不得聲稱工具執行了實際未曾完成的工作。

Examples of AI-Assisted Manuscript Preparation / 人工智慧輔助稿件準備的示例

  • Language Editing and Grammar Correction: Assistance with spelling, grammar, language expression, translation, or the refinement of existing manuscript text. Authors should ensure that such assistance preserves the intended scholarly meaning, disciplinary terminology, and accuracy of the original material.

    語言編輯與文法修正: 協助拼寫、文法、語言表達、翻譯,或對稿件既有文字進行潤飾。作者應確保相關協助保留原有學術意義、學科術語及材料準確性。

  • Readability and Presentation: Assistance with sentence clarity, organization, readability, or formatting without misrepresenting the underlying scholarly content. Structural suggestions generated by AI should be evaluated by the authors rather than adopted without review.

    可讀性與內容呈現: 協助改善句子清晰度、內容組織、可讀性或排版,但不得因此錯誤呈現原始學術內容。人工智慧提出的結構建議應由作者評估,而非未經檢查便直接採用。

  • Limited Drafting Assistance: Assistance with preparing or revising portions of text under meaningful human supervision, where authors independently verify the content, assess its intellectual appropriateness, identify relevant sources, and retain responsibility for the final manuscript.

    有限的文本起草協助: 在具有實質意義的人類監督下協助起草或修訂部分文字,並由作者獨立核實內容、評估其智識適切性、確認相關來源,以及對最終稿件負責。

Authors should include an AI-use statement that accurately identifies the assistance received. Where applicable and available, the statement should indicate the tool or system used, its relevant version, the purpose of its use, and the extent to which it contributed to preparing the manuscript.

作者應提供準確描述所接受協助的人工智慧使用聲明。在適用且資訊可取得的情況下,聲明應包括所使用的工具或系統、相關版本、使用目的,以及其對稿件準備工作的參與程度。

The statement should be included in an appropriate part of the manuscript or the submission information, consistent with the nature of the use and the journal's applicable submission instructions. Assistance with writing or language preparation should be distinguished from AI technologies used as substantive research methods, which must also be described in the relevant methodological discussion.

使用聲明應依據實際用途及本刊適用投稿指引,置於稿件或投稿資訊中的適當位置。文字起草或語言準備方面的協助,應與作為實質研究方法使用的人工智慧技術加以區分;後者亦須在相關方法論說明中得到交代。

An appropriate disclosure may explain that a tool was used for language refinement, limited text drafting, translation, or another identifiable task, while specifying that the authors reviewed and verified the resulting content. The statement should reflect the actual use of the technology rather than rely on vague or misleading wording.

適當的披露可以說明工具曾用於語言潤飾、有限的文字起草、翻譯或其他可以辨識的工作,並指出作者已檢查與核實相關內容。聲明應反映技術的實際使用情況,而非採用模糊或可能產生誤導的表述。

Illustrative AI-Use Disclosure / 人工智慧使用聲明示例

During manuscript preparation, the authors used an AI-assisted language-editing application to improve grammar, sentence clarity, and readability. The authors reviewed and verified the resulting text, checked the accuracy of relevant claims and references, and remain responsible for the originality, accuracy, and integrity of the final manuscript.

在稿件準備過程中,作者使用人工智慧輔助語言編輯應用,以改善文法、句子清晰度及可讀性。作者已檢查並核實相關文字、確認重要主張與參考文獻的準確性,並對最終稿件的原創性、準確性及完整性承擔責任。

This example illustrates a limited language-editing use and is not a mandatory wording template. Authors must adapt their statements to the actual tools and activities involved, including identifying the tool and relevant version where applicable and available. The example should not be used to describe research analysis, image generation, data processing, or substantial drafting when those activities formed part of the actual use.

上述示例僅說明有限的語言編輯用途,並非強制使用的固定聲明格式。作者必須根據實際工具與使用活動調整說明,並在適用且資訊可取得的情況下標示工具及相關版本。如實際使用涉及研究分析、圖像生成、資料處理或大量文字起草,便不應以此語言編輯示例取代相關用途的準確披露。

Disclosure does not replace appropriate citation, methodological reporting, verification of factual statements, or the obligation to obtain necessary permissions. A complete disclosure should make the use of AI understandable without implying that the system independently verified scholarly claims or assumed responsibility for the manuscript.

披露人工智慧使用情況,不能取代適當引用、方法論報告、事實核實或取得必要許可的義務。完整的使用聲明應使相關用途容易理解,同時避免暗示人工智慧系統已獨立核實學術主張或承擔稿件責任。

Where AI-assisted technologies have materially influenced the creation or modification of figures, images, tables, or other scholarly content, authors should explain the nature of that involvement and ensure that the resulting materials are accurately represented. The relevant description may belong in a methodological explanation, figure legend, or another appropriate part of the manuscript, depending on the role of the technology.

如人工智慧輔助技術實質影響圖表、圖像、表格或其他學術內容的製作或修改,作者應說明相關參與的性質,並確保最終材料得到準確呈現。視技術所扮演的角色,相關說明可以置於方法論、圖說或稿件其他適當部分。

Author Responsibility / 作者責任

Authors retain full responsibility for the scholarly content of their manuscripts, including text, analyses, interpretations, references, and other materials prepared with AI assistance. Responsibility cannot be transferred to an AI provider, model, software application, or automated system.

作者須對稿件的全部學術內容承擔責任,包括經人工智慧協助準備的文字、分析、解釋、參考文獻及其他材料。相關責任不能轉移予人工智慧服務提供者、模型、軟體應用或自動化系統。

Human responsibility requires more than reading AI-generated output for fluency. Authors must evaluate whether the information is accurate, whether cited sources genuinely support the stated claims, whether interpretations are justified by the relevant evidence, and whether the resulting material preserves the originality and integrity of the work.

人類作者的責任不僅在於檢查人工智慧生成文字是否流暢,亦須評估資訊是否準確、所引用來源是否真正支持相關主張、研究解釋是否具有充分證據依據,以及最終材料是否維持作品的原創性與完整性。

Responsibilities That Remain with Authors / 作者不得轉移的責任

  • Accuracy and Scholarly Validity: Verify factual statements, analytical claims, methodological descriptions, and conclusions before including them in the manuscript. Where the validity of a statement cannot be adequately established, the statement should be corrected, qualified, or removed.

    準確性與學術有效性: 在將內容納入稿件前,核實事實陳述、分析主張、方法論描述及研究結論。對無法獲得充分支持的陳述,應予以修正、適當限定或刪除。

  • Originality and Appropriate Attribution: Ensure that AI-assisted material does not reproduce or misappropriate the intellectual contributions of others without appropriate acknowledgment or permission. Authors must distinguish independently developed analysis from material derived from existing scholarship.

    原創性與適當來源標示: 確保人工智慧輔助內容不會在缺乏適當確認或許可的情況下,重製或不當挪用他人的智識貢獻。作者必須區分獨立形成的分析與建立在既有研究成果之上的內容。

  • Verification of References: Confirm that cited publications, authors, titles, publication details, quotations, and identifiers correspond to real and accurately represented sources. References generated or suggested by AI must be checked against the original publications or reliable bibliographic records.

    參考文獻核實: 確認所引用的出版成果、作者、題名、出版資訊、引文及識別碼均對應真實且被準確呈現的來源。由人工智慧生成或建議的參考文獻,必須與原始出版成果或可靠的書目紀錄核對。

  • Integrity of Research Findings: Prevent the inclusion of fabricated data, invented observations, unsupported results, or misleading representations of research that was not actually performed. AI-generated simulations or synthetic materials must be distinguished from genuine empirical observations.

    研究發現的完整性: 防止稿件包含偽造資料、虛構觀察、缺乏證據的結果,或對實際並未執行的研究作出誤導性描述。人工智慧生成的模擬或合成材料,必須與真實實證觀察加以區分。

  • Protection of Confidential Information: Ensure that the use of AI tools does not involve unauthorized disclosure of personal, sensitive, confidential, proprietary, or otherwise protected research materials. Applicable consent, privacy, contractual, and institutional obligations must be respected.

    保密資訊保護: 確保人工智慧工具的使用不涉及未經授權披露個人、敏感、保密、專有或其他受保護的研究材料,並遵守適用的同意、隱私、契約及機構義務。

  • Human Review and Accountability: Exercise meaningful independent judgment over AI-assisted material, make informed decisions concerning its inclusion, and accept responsibility for the manuscript submitted for editorial evaluation. Human review should address the substance of the work, not merely its wording or presentation.

    人類審查與責任承擔: 對人工智慧輔助內容進行具有實質意義的獨立判斷、就其是否納入稿件作出知情決定,並對提交編輯評估的稿件承擔責任。人類審查應涉及學術內容本身,而非僅檢查文字或呈現方式。

Authors should not assume that fluent language, plausible citations, or technically sophisticated output establishes the reliability of AI-generated content. AI systems may produce inaccurate descriptions, nonexistent references, fabricated quotations, incomplete explanations, biased representations, or interpretations unsupported by the original evidence.

作者不應認為語言流暢、引用看似可信,或輸出具有技術複雜性,便代表人工智慧生成的內容必然可靠。人工智慧系統可能產生不準確的描述、不存在的參考文獻、虛構引文、不完整的解釋、帶有偏差的呈現,或缺乏原始證據支持的研究判斷。

The verification of AI-assisted literature summaries requires particular care. A generated summary may omit important qualifications, conflate findings from different sources, attribute claims to the wrong publication, or present an interpretation that is stronger than the evidence permits. Authors should therefore consult the underlying literature when evaluating claims that are important to their argument.

人工智慧輔助文獻摘要的核實尤其需要審慎處理。生成的摘要可能省略重要限制、混淆不同來源的研究發現、將主張錯誤歸於某項出版成果,或提出超出證據所能支持範圍的解釋。因此,作者在評估對論證具有重要性的主張時,應查閱相關原始文獻。

Authors are expected to examine material generated or modified by AI tools before submission and to correct or remove content that cannot be substantiated. The journal's requirements concerning plagiarism, research data, publication ethics, and intellectual contribution apply regardless of whether AI assistance was involved.

作者應在投稿前檢查由人工智慧生成或修改的內容,並修正或刪除無法獲得支持的材料。無論是否使用人工智慧輔助,本刊關於剽竊、研究資料、出版倫理及智識貢獻的要求均同樣適用。

Authors should also ensure that AI-assisted changes do not introduce discrepancies between the manuscript and the research actually performed. Language editing, restructuring, or summarization must not change numerical results, methodological descriptions, participant characteristics, quotations, or the substantive meaning of findings without appropriate verification and correction.

作者亦應確保人工智慧輔助修改不會使稿件內容與實際執行的研究產生不一致。語言編輯、結構重組或摘要處理,不得在缺乏適當核實與修正的情況下,改變數值結果、方法論描述、研究參與者特徵、引文或研究發現的實質意義。

AI in Research Methods / 研究方法中的人工智慧

When AI technologies constitute part of the research methodology, authors should describe their use in the relevant methodological section of the manuscript. Such applications may include data analysis, computational modeling, classification, automated text processing, image analysis, simulation, or other tasks in which AI systems materially influence the research process or its findings.

當人工智慧技術構成研究方法的一部分時,作者應在稿件的相關方法論章節說明其使用情況。這些應用可能包括資料分析、計算建模、分類、自動化文本處理、圖像分析、模擬,或其他由人工智慧系統實質影響研究過程或研究發現的工作。

Methodological transparency is necessary when AI systems influence the production, selection, transformation, classification, or interpretation of research evidence. Authors should explain the role of the technology sufficiently for readers and reviewers to assess the appropriateness of the methods, understand the origin of relevant outputs, and identify limitations affecting the conclusions.

當人工智慧系統影響研究證據的產生、選擇、轉換、分類或解釋時,方法論透明度具有重要性。作者應充分說明技術的作用,使讀者與審稿人能夠評估方法是否適當、理解相關輸出的來源,並辨識影響研究結論的局限。

Methodological Information / 方法論說明事項

  • System Identification: Identify the AI system, software, model, or tool used, including the relevant version when available and necessary for understanding the research. Where the system was accessed through a service or platform, relevant information about that environment may also assist interpretation.

    系統識別: 說明所使用的人工智慧系統、軟體、模型或工具,並在可取得且有助於理解研究時註明相關版本。如透過特定服務或平台使用系統,相關環境資訊亦可能有助於研究解釋。

  • Research Purpose: Explain the function performed by the AI system and why its use was appropriate to the research question or methodological approach. Authors should distinguish supporting technical operations from analyses or decisions that materially influence the research outcomes.

    研究目的: 解釋人工智慧系統執行的工作,以及其使用為何符合研究問題或方法論路徑。作者應區分輔助性技術操作與實質影響研究結果的分析或決策。

  • Procedures and Relevant Settings: Describe the inputs, processing steps, configurations, prompts, parameters, or other methodological details necessary to understand or appropriately evaluate the research process, subject to legitimate confidentiality and intellectual property restrictions.

    程序與相關設定: 說明理解或適當評估研究過程所需的輸入內容、處理步驟、系統設定、提示詞、參數或其他方法論細節,但相關說明仍須遵守正當的保密與智慧財產權限制。

  • Human Oversight and Validation: Explain how authors assessed the system's output, including relevant checks of accuracy, consistency, reliability, or suitability for the intended research purpose. Where appropriate, this may involve comparison with independently established evidence, manual review, or other justified validation procedures.

    人類監督與結果驗證: 說明作者如何評估系統輸出,包括對準確性、一致性、可靠性或研究目的適切性的相關檢查。在適當情況下,可透過與獨立確立的證據比較、人工審查或其他具有合理依據的驗證程序進行評估。

  • Limitations and Interpretive Caution: Identify material limitations, uncertainties, sources of potential error, or biases that may affect the interpretation of AI-assisted results. The implications of these limitations should be considered when determining the strength and scope of the manuscript's conclusions.

    局限與解釋上的審慎性: 辨識可能影響人工智慧輔助結果解釋的重要限制、不確定性、潛在誤差來源或偏差。判斷稿件結論的強度及適用範圍時,應考慮相關局限的影響。

Methodological reporting should be proportionate to the role played by the technology. A tool used for a minor supporting operation may require a different level of detail from a model whose outputs form a substantial part of the research evidence or analytical framework. The objective is sufficient transparency for scholarly evaluation, not the indiscriminate reproduction of every technical interaction.

方法論報告的詳細程度應與技術在研究中所扮演的角色相稱。用於次要輔助工作的工具,與其輸出構成研究證據或分析框架重要部分的模型,可能需要不同程度的說明。其目的在於提供足以支持學術評估的透明度,而非無差別地重現所有技術互動。

Where a model's outputs are sensitive to prompts, configurations, software versions, or stochastic variation, these factors may be relevant to evaluating the reliability and reproducibility of the results. Authors should describe material sources of variability and avoid claiming that an analysis is fully reproducible merely because a tool or model name has been supplied.

如模型輸出容易受到提示詞、設定、軟體版本或隨機性變化的影響,這些因素便可能與結果的可靠性及可重現性有關。作者應說明具有實質影響的變異來源,並避免僅因已提供工具或模型名稱,便宣稱相關分析可以完全重現。

In studies involving AI-based classification, coding, translation, transcription, or interpretation, authors should consider whether the system's errors or biases could affect the research findings. Validation should be appropriate to the actual research objective and should not be represented as more comprehensive than the procedures performed.

當研究涉及人工智慧分類、編碼、翻譯、轉錄或詮釋時,作者應考慮系統錯誤或偏差是否可能影響研究發現。驗證方式應與實際研究目的相符,且不得將已執行的驗證程序描述為比實際情況更加全面。

AI-Generated and Synthetic Research Materials / 人工智慧生成與合成研究材料

AI-generated or synthetic material may be a legitimate component of certain research designs, including simulation, methodological testing, the evaluation of AI systems, or investigations that explicitly examine generated content. Where such material is used, authors must accurately identify it as generated or synthetic and explain its role in the study.

人工智慧生成或合成材料可以構成某些研究設計的正當組成部分,包括模擬研究、方法測試、人工智慧系統評估,或明確以生成內容為研究對象的調查。使用此類材料時,作者必須準確標示其生成或合成性質,並說明其在研究中的作用。

Synthetic data must not be misrepresented as observations collected from actual participants, real-world events, or experiments that were not conducted. Similarly, AI-generated quotations, reconstructed dialogues, images, or illustrative examples must not be presented as authentic records of persons, events, or research procedures when they are not.

合成資料不得被錯誤呈現為從真實研究參與者、現實事件或實際未執行之實驗中取得的觀察結果。同樣,人工智慧生成的引文、重構對話、圖像或示例,如並非真實人物、事件或研究程序的紀錄,便不得被呈現為具有真實紀錄性質的材料。

Disclosure of synthetic material is necessary, but it does not independently establish that the research design, analysis, or conclusions are methodologically valid. Authors should distinguish findings derived from real-world observations from results produced through simulation or generation and explain the limitations of using synthetic evidence to draw conclusions about actual phenomena.

披露合成材料的使用具有必要性,但披露本身並不能獨立證明研究設計、分析或結論在方法論上有效。作者應區分建立在現實觀察之上的研究發現與透過模擬或生成得到的結果,並解釋使用合成證據對真實現象作出推論所涉及的局限。

When research involves confidential datasets, human participants, proprietary materials, or other protected information, authors must ensure that AI-assisted processing is compatible with the relevant consent, privacy, ethical, legal, and contractual requirements. Data protection obligations apply regardless of whether the AI system is used for analysis, generation, classification, translation, or another research task.

如研究涉及保密資料集、人類研究參與者、專有材料或其他受保護資訊,作者必須確保人工智慧輔助處理符合相關同意、隱私、倫理、法律及契約要求。無論人工智慧系統用於分析、生成、分類、翻譯或其他研究工作,資料保護義務均同樣適用。

The use of AI in a research method does not automatically establish the reliability of that method or the validity of its results. AI-assisted procedures should be evaluated according to the same principles of methodological appropriateness, evidence quality, research transparency, and proportional interpretation that apply to other scholarly methods.

在研究方法中使用人工智慧,並不自動證明相關方法可靠或結果有效。人工智慧輔助程序仍應依據適用於其他學術方法的原則,接受方法論適切性、證據品質、研究透明度及結論解釋相稱性的評估。

Prohibited Uses of AI / 禁止的人工智慧用途

Silence does not permit the use of AI technologies in ways that fabricate or misrepresent research, conceal material information about the scholarly process, undermine authorship accountability, violate confidentiality, or manipulate editorial and peer-review procedures.

《Silence》不允許以人工智慧技術偽造或錯誤呈現研究、隱瞞影響學術程序的重要資訊、削弱作者責任、違反保密義務,或操縱編輯與同行評審程序。

These restrictions concern the substance and consequences of the conduct, not merely the technical characteristics of a particular tool. The use of AI is unacceptable when it creates false scholarly evidence, misrepresents authorship or research activities, compromises protected information, or interferes improperly with editorial evaluation.

上述限制針對相關行為的實質內容與後果,而非僅依據特定工具的技術特徵判斷。當人工智慧使用造成虛假學術證據、錯誤呈現作者資格或研究活動、損害受保護資訊,或不當干預編輯評估時,即屬不可接受的行為。

Unacceptable Practices / 不可接受的行為

  • Undisclosed Generation or Misrepresentation of Research Evidence: Presenting AI-generated data, observations, analyses, images, or results as independently obtained empirical evidence without accurately disclosing their origin and methodological status. This includes portraying synthetic observations as records from actual participants, experiments, or events that did not occur.

    未披露生成內容或錯誤呈現研究證據: 未準確披露人工智慧生成之資料、觀察、分析、圖像或結果的來源與方法論性質,便將其呈現為獨立取得的實證證據,包括將合成觀察錯誤描述為來自實際未曾發生的參與者活動、實驗或事件。

  • Fabrication of Scholarly Sources or Findings: Creating or submitting nonexistent references, fabricated citations, invented quotations, false observations, altered research records, or findings that are not supported by actual evidence. Apparently plausible AI output does not constitute a legitimate source when the underlying publication or observation cannot be verified.

    偽造學術來源或研究發現: 製作或提交不存在的參考文獻、虛構引用、捏造引文、虛假觀察、經不當修改的研究紀錄,或缺乏真實證據支持的研究發現。當底層出版成果或觀察無法核實時,外觀可信的人工智慧輸出並不構成正當學術來源。

  • Unsupervised Substantial Text Generation: Submitting substantial AI-generated manuscript content without meaningful human examination, verification, intellectual oversight, or acceptance of responsibility. Authors must not use automated text generation to conceal the absence of independent scholarly analysis or to misrepresent the origin of substantive intellectual contributions.

    缺乏監督的大量文本生成: 在缺乏具有實質意義的人類檢查、核實、智識監督或責任承擔的情況下,提交大量人工智慧生成的稿件內容。作者不得利用自動化文本生成掩蓋缺乏獨立學術分析的事實,或錯誤呈現實質智識貢獻的來源。

  • Manipulation of Editorial or Peer-Review Procedures: Using AI systems to fabricate reviewer identities, create deceptive communications, generate misleading assessments, impersonate individuals, or otherwise interfere improperly with editorial decision-making or the integrity of peer review.

    操縱編輯或同行評審程序: 使用人工智慧系統偽造審稿人身份、製作欺騙性通信、生成具有誤導性的評估、冒充他人,或以其他不當方式干預編輯決策與同行評審的完整性。

  • Unauthorized Disclosure of Protected Materials: Entering confidential, sensitive, personal, proprietary, unpublished, or otherwise protected research information into AI services where such processing is not appropriately authorized. The convenience of an external AI platform does not remove obligations relating to privacy, consent, confidentiality, or intellectual property.

    未經授權披露受保護材料: 在未取得適當授權的情況下,將保密、敏感、個人、專有、未發表或其他受保護的研究資訊提供予人工智慧服務處理。外部人工智慧平台的便利性,不會取消隱私、知情同意、保密或智慧財產權方面的義務。

The distinction between legitimate AI-supported research and unacceptable fabrication depends on accurate reporting, methodological validity, and the nature of the evidence. Clearly labeled synthetic data used in an appropriate simulation or validation study are not equivalent to invented empirical observations falsely attributed to real participants or experiments.

正當的人工智慧輔助研究與不可接受的資料偽造之間,應根據報告準確性、方法論有效性及證據性質加以區分。在適當模擬或驗證研究中使用、且已明確標示的合成資料,並不等同於被錯誤歸於真實參與者或實驗的虛構實證觀察。

Similarly, a disclosure statement cannot make fabricated evidence, plagiarism, misleading authorship, or unauthorized disclosure ethically acceptable. Authors must satisfy the journal's substantive research and publication standards in addition to reporting their use of AI tools.

同樣,人工智慧使用聲明不能使偽造證據、剽竊、不當作者署名或未經授權的資訊披露在倫理上變得可以接受。作者除披露人工智慧使用情況外,仍須符合本刊實質性的研究與出版標準。

The significance of an AI-related concern should be assessed in context. Incomplete disclosure, factual inaccuracy, fabricated evidence, and deliberate manipulation are not necessarily equivalent. Editorial consideration should distinguish correctable omissions from serious conduct that undermines the reliability or integrity of the submitted work.

涉及人工智慧使用的問題應結合具體情況評估。披露不完整、事實不準確、偽造證據及蓄意操縱,並不必然具有相同的性質或嚴重程度。編輯評估應區分可以修正的遺漏,以及實質損害投稿可靠性或完整性的重大行為。

Concerns involving prohibited practices may be examined under the journal's Publication Ethics Policy. Depending on the available evidence and the seriousness of the issue, the editorial office may request clarification, seek revision, reject a manuscript, or take another appropriate action consistent with applicable procedures. Suspected misconduct should not be treated as an established finding before appropriate assessment.

涉及禁止性行為的疑慮,可依照本刊 Publication Ethics Policy(出版倫理政策)進行審查。編輯部可根據現有證據及問題嚴重程度,要求澄清、提出修訂要求、拒絕稿件,或依適用程序採取其他適當措施。在完成適當評估之前,不應將涉嫌學術不端直接視為已確認的事實。

Editorial Use of AI / 編輯與審稿過程中的人工智慧使用

Editors and reviewers may use limited AI-assisted technologies for appropriate supporting tasks, such as language checking or administrative work, provided that their use is compatible with confidentiality, research integrity, information security, and the journal's editorial responsibilities. Such use does not authorize the disclosure of unpublished manuscripts or confidential editorial information to external systems.

編輯與審稿人可以在適當情況下使用有限的人工智慧輔助技術處理語言檢查或行政工作,但相關使用必須符合保密、研究誠信、資訊安全及本刊編輯責任等要求。此類使用並不構成向外部系統披露未發表稿件或保密編輯資訊的授權。

AI-assisted editorial tasks must remain subordinate to human professional judgment. The use of an automated system does not transfer responsibility for confidentiality, the accuracy of editorial communication, the evaluation of scholarship, or the fairness of decisions away from the editor or reviewer involved.

人工智慧輔助編輯工作必須服從人類的專業判斷。使用自動化系統不會使相關編輯或審稿人免於承擔保密、編輯通信準確性、學術評估及決策公正性方面的責任。

Confidentiality of Submitted Manuscripts / 投稿稿件的保密性

Submitted manuscripts, supplementary materials, reviewer reports, editorial correspondence, and unpublished research information must be handled as confidential materials where applicable. Editors and reviewers must not upload or disclose such content to unauthorized external AI services or other third-party systems that could compromise the journal's confidentiality obligations.

投稿稿件、補充材料、審稿報告、編輯通信及未公開研究資訊,在適用情況下均須作為保密材料處理。編輯與審稿人不得將相關內容上傳或披露予未獲授權的外部人工智慧服務或其他可能損害本刊保密義務的第三方系統。

The use of an AI tool for a seemingly limited task, including language correction, translation, summarization, or document organization, may still involve processing confidential manuscript content. The nature of the information transmitted, the authorization for its use, the conditions imposed by the service provider, and applicable data protection requirements must therefore be considered before employing such a service.

即使人工智慧工具僅用於語言修正、翻譯、內容摘要或文件整理等看似有限的工作,也可能涉及保密稿件內容的處理。因此,在使用相關服務前,仍須考慮所傳輸資訊的性質、使用授權、服務提供者所設定的條件,以及適用的資料保護要求。

Reviewers must not independently disclose or upload unpublished manuscripts, substantial excerpts, research data, figures, or confidential review communications to external AI services without appropriate editorial authorization and the safeguards necessary to preserve confidentiality. Where the use of a service could affect authors' rights or the confidentiality of submitted material, authorization and the relevant protections must be considered before any disclosure takes place.

未經適當編輯授權,且尚未確認維持保密所需的保障措施之前,審稿人不得自行將未公開稿件、實質性內容片段、研究資料、圖像或保密評審通信披露或上傳至外部人工智慧服務。如相關服務的使用可能影響作者權利或投稿材料的保密性,便必須在任何披露發生之前考慮必要授權及保護措施。

Authorization to use a particular tool should not be inferred merely because the tool is commercially available or routinely employed for other academic tasks. A system that is appropriate for editing publicly available text may not be suitable for processing unpublished manuscripts, identifiable participant information, or confidential reviewer comments.

不能僅因某項工具可在市場上取得,或經常用於其他學術工作,便推定其已獲授權使用。適合編輯公開文字的系統,未必適合處理未發表稿件、可識別研究參與者的資訊,或保密審稿意見。

Human Editorial Judgment / 人類編輯判斷

Editorial decisions are made by human editors on the basis of scholarly assessment, relevant reviewer recommendations, and the journal's applicable editorial and ethical requirements. AI-generated evaluations or recommendations must not be treated as substitutes for independent professional judgment.

編輯決定須由人類編輯根據學術評估、相關審稿建議,以及本刊適用的編輯與倫理要求作出。人工智慧生成的評價或建議不得取代獨立的專業判斷。

The assessment of originality, scholarly contribution, methodological adequacy, research ethics, and publication suitability requires accountable editorial consideration. Automated classification, language analysis, or other technical assistance may inform limited aspects of a workflow but cannot independently determine whether a manuscript satisfies the journal's scholarly standards.

對原創性、學術貢獻、方法論充分性、研究倫理及出版適切性的評估,必須由能夠承擔責任的編輯進行審慎判斷。自動化分類、語言分析或其他技術協助可以支持工作流程中的有限環節,但不能獨立決定稿件是否達到本刊學術標準。

Reviewers remain responsible for the substance of their assessments. Any permitted technical assistance must not undermine the independence, accuracy, confidentiality, or accountability of a review report. Reviewers should disclose relevant permitted AI assistance to the editor, and the responsibilities set out in the journal's Reviewer Guidelines and Peer Review Policy continue to apply.

審稿人仍須對自身評估的實質內容負責。任何獲允許的技術協助均不得削弱評審報告的獨立性、準確性、保密性或責任承擔。審稿人應向編輯披露相關獲准使用的人工智慧協助,而本刊 Reviewer Guidelines(審稿人指南)及 Peer Review Policy(同行評審政策)所規定的責任仍然適用。

Editors should also distinguish automated text analysis from a substantiated finding of academic misconduct. AI-detection systems and other automated assessments may have important limitations and should not be treated as conclusive evidence that a manuscript was generated by AI or that an author violated the journal's policies. Where concerns arise, they should be examined through appropriate editorial assessment and relevant evidence.

編輯亦應區分自動化文字分析與已獲充分證據支持的學術不端判定。人工智慧內容檢測系統及其他自動化評估可能存在重要限制,不應被直接視為足以證明稿件由人工智慧生成,或作者違反期刊政策的決定性證據。如出現相關疑慮,應透過適當的編輯評估及相關證據加以審查。

Protection of Unpublished Intellectual Contributions / 未發表智識貢獻的保護

AI-assisted editorial work must not result in unauthorized disclosure, appropriation, or reuse of unpublished ideas, data, methods, interpretations, or other intellectual contributions obtained through manuscript evaluation. Confidentiality obligations continue after an editorial decision has been made.

人工智慧輔助編輯工作不得導致在稿件評估過程中取得的未發表思想、資料、方法、研究解釋或其他智識貢獻遭到未經授權的披露、挪用或再利用。編輯決定作出後,相關保密義務仍然持續。

The protection of unpublished intellectual contributions applies regardless of whether a manuscript is accepted, revised, or rejected. Editors and reviewers must not use confidential material obtained through the review process for independent research, professional advantage, or other unauthorized purposes. Any permitted technical assistance must preserve these protections.

無論稿件最終獲錄用、進入修訂或被拒絕,未發表智識貢獻仍應受到保護。編輯與審稿人不得將評審過程中取得的保密材料用於個人研究、職業利益或其他未獲授權的目的。任何獲准使用的技術協助均須維持相關保護。

Where AI is used in a permitted editorial or production activity, the individuals responsible for that activity remain accountable for the accuracy of the resulting work and the protection of confidential information. The involvement of a technical service does not change the scholarly responsibilities associated with the journal's editorial decisions or the rights of authors whose manuscripts are under consideration.

如在獲允許的編輯或出版製作活動中使用人工智慧,負責相關工作的個人仍須對結果的準確性及保密資訊的保護承擔責任。技術服務的參與不會改變期刊編輯決策所涉及的學術責任,也不會改變接受評估之稿件作者所享有的權利。

Commitment to Responsible Use of Technology / 對負責任使用技術的承諾

Silence supports the responsible, transparent, and academically justified use of emerging technologies in scholarly communication. The journal recognizes that AI may provide useful assistance in research and manuscript preparation, while also introducing challenges involving accuracy, attribution, methodological transparency, confidentiality, and intellectual accountability.

《Silence》支持在學術交流中負責任、透明且具有合理學術依據地使用新興技術。本刊認為,人工智慧可以為研究與稿件準備提供有價值的協助,但同時也可能帶來準確性、來源標示、方法論透明度、保密及智識責任等方面的挑戰。

Responsible AI use requires more than disclosure alone. Authors must understand and evaluate the role of the systems they employ, verify material claims, accurately describe relevant research procedures, and protect the rights and interests of individuals and organizations whose information may be involved.

負責任的人工智慧使用不僅需要披露。作者亦須理解並評估所使用系統的作用、核實重要主張、準確描述相關研究程序,並保護可能涉及之個人及機構的權利與利益。

The journal's openness to unconventional scholarly contributions, negative findings, unresolved research questions, and interdisciplinary methods does not diminish these requirements. AI-assisted research is evaluated according to the quality, integrity, and evidentiary basis of its scholarly contribution, rather than the mere presence or absence of AI technology.

本刊對非傳統學術成果、陰性發現、未決研究問題及跨學科方法的開放態度,不會降低上述要求。人工智慧輔助研究應根據其學術貢獻的品質、完整性及證據基礎接受評估,而非僅以是否使用人工智慧技術作為判斷依據。

The responsible use of technology also requires recognition of its limitations. AI systems may be appropriate for certain forms of language assistance, data processing, computational modeling, or exploratory analysis while remaining unsuitable as independent authorities on factual accuracy, scholarly originality, ethical acceptability, or editorial merit. Those judgments require attention to the evidence and circumstances of the particular work.

負責任地使用技術,也需要認識其局限。人工智慧系統可能適合提供某些語言協助、資料處理、計算建模或探索性分析,但不適合作為事實準確性、學術原創性、倫理適切性或編輯價值的獨立權威。相關判斷仍須結合個別作品的證據與具體情況進行。

Transparency should make the scholarly process more understandable rather than create a superficial appearance of compliance. An AI-use statement is meaningful when it accurately identifies the technology's role, while methodological reporting is meaningful when it explains how research outputs were generated, evaluated, and interpreted. Both should support readers' ability to assess the contribution without overstating what has been verified.

透明度應使學術程序更加容易理解,而非僅形成表面上的合規形象。當人工智慧使用聲明準確辨識技術的角色時,披露才具有實質意義;當方法論報告解釋研究輸出如何產生、接受評估及受到解釋時,相關說明才能真正支持學術檢驗。兩者均應協助讀者評估研究貢獻,同時避免誇大已完成的驗證程度。

The integrity of scholarly publishing depends on a clear allocation of responsibility. Authors remain accountable for the research they submit, reviewers for the independent assessments they provide, and editors for the decisions made through the journal's editorial process. AI-assisted technologies may support aspects of these activities but do not replace the corresponding human responsibilities.

學術出版的完整性依賴清晰的責任分配。作者須對所提交的研究負責,審稿人須對其提供的獨立評估負責,而編輯則須對期刊編輯程序中作出的決策負責。人工智慧輔助技術可以支持上述工作的部分環節,但不能取代相應的人類責任。

By emphasizing transparency, human oversight, appropriate attribution, methodological rigor, and responsible scholarly judgment, Silence aims to support technological innovation without compromising the reliability of the academic record or the ethical responsibilities of authors, editors, and reviewers.

透過重視透明度、人類監督、適當來源標示、方法論嚴謹性及負責任的學術判斷,《Silence》期望在支持技術創新的同時,維護學術記錄的可靠性,以及作者、編輯與審稿人所應承擔的倫理責任。

Integrity · Transparency · Responsibility

誠信 · 透明 · 責任