AI & Scholarship

At AI & Antiquity, artificial intelligence is neither treated as a substitute for scholarship nor approached as a technology that should be excluded from academic practice. We understand AI as both an emerging object of scholarly inquiry and a potentially valuable set of tools for research, teaching, accessibility, and academic communication. Its value, however, depends on how it is used. Artificial intelligence can support intellectual work, but it cannot replace disciplinary expertise, critical interpretation, scholarly judgement, or human accountability.

Our position is therefore neither celebratory nor prohibitive. We support the critical, transparent, verifiable, proportionate, and ethically responsible use of AI within Ancient Studies. The central question is not simply whether artificial intelligence has been used, but what role it has played, what assumptions and limitations accompany that use, what costs it entails, and whether it contributes meaningfully to rigorous scholarship, inclusive education, or the communication of knowledge.

Our Position

 

AI as a Tool

Artificial intelligence can assist with activities such as language revision, translation, information organisation, coding, data analysis, the creation of teaching materials, accessibility support, and exploratory research. Used carefully, these tools can reduce repetitive work, facilitate multimodal learning, and open new ways of interacting with complex historical and archaeological materials.

AI as an Object of Inquiry

AI systems also deserve critical examination in their own right. Their representations of Antiquity are shaped by training data, technical design, linguistic conventions, cultural assumptions, commercial infrastructures, and contemporary narratives about the ancient world. Studying what AI produces — and what it omits, simplifies, distorts, or invents — can itself become a productive form of historical, pedagogical, and digital inquiry.

Artificial intelligence may assist scholarship, but scholarly responsibility remains human. Authors, teachers, editors, reviewers, and researchers remain accountable for the arguments they make, the evidence they use, the interpretations they advance, the decisions they take, and the accuracy and integrity of the materials they produce.

Principles for Responsible AI Use

 

Transparency

Substantial use of generative AI should be disclosed clearly enough for readers, reviewers, and editors to understand the role played by the technology.

Verification

AI-generated claims, references, translations, interpretations, code, and visual materials must be checked against appropriate scholarly evidence and reliable sources.

Human Accountability

AI systems cannot assume authorship or scholarly responsibility. Intellectual judgement and responsibility for published work remain exclusively human.

Critical Awareness

Users should remain attentive to bias, hallucination, homogenisation, cultural assumptions, loss of nuance, unequal access, and the technical limitations of generative systems.

For the purposes of disclosure, substantial use means AI assistance that materially contributes to the generation, analysis, interpretation, organisation, or presentation of scholarly content. Routine functions such as spelling correction, basic formatting, reference-management automation, or standard software features do not normally require disclosure unless they materially affect the intellectual content of the work.

Pluralism, Proportionality & Human Value

 

Our support for the responsible use of artificial intelligence does not imply that scholars, educators, students, authors, editors, or reviewers should be expected to use it. AI & Antiquity fully respects positions ranging from active experimentation to principled non-use. Decisions not to use artificial intelligence may arise from ethical or moral concerns, environmental considerations, intellectual convictions, professional judgement, privacy requirements, pedagogical principles, or simply a preference for forms of work in which human practice itself is considered valuable. Such choices deserve the same scholarly respect as responsible decisions to employ AI.

Artificial intelligence is a rapidly evolving and genuinely contested area of academic practice. Questions concerning its ethical legitimacy, environmental impact, epistemic consequences, pedagogical value, labour implications, and appropriate place within scholarship remain open to serious debate. AI & Antiquity does not seek to impose a single institutional position on questions for which reasonable disagreement persists. Members of our Editorial Team, Advisory Board, contributors, and academic colleagues may — and sometimes do — disagree openly about when AI should be used, whether particular uses are desirable, or whether some applications should be rejected altogether. We regard such disagreement not as a weakness to be concealed but as an essential feature of responsible scholarly debate.

Nothing in this statement should be interpreted as requiring authors, reviewers, editors, members of the Editorial or Advisory Boards, or collaborating scholars to endorse artificial intelligence or to share a common assessment of its desirability. Participation in the journal does not imply agreement with every position expressed in this statement beyond the procedural standards necessary for ethical publishing. Critical, sceptical, restrictive, and non-use positions are as legitimate within the journal’s intellectual community as arguments in favour of responsible experimentation. The journal is committed to providing space for these competing perspectives when they are expressed through evidence, reasoned argument, intellectual honesty, and mutual respect.

The Freedom Not to Use AI

No author, reviewer, editor, teacher, or student should be treated as technologically resistant, professionally deficient, or intellectually conservative merely because they choose not to use artificial intelligence. Non-use may itself be a thoughtful and ethically informed position. The journal therefore rejects the assumption that technological adoption is inherently progressive or that every academic practice must be reorganised around emerging tools.

Proportionality

AI should not be used merely because it is available. Its use should be proportionate to the problem being addressed and to the value it actually adds. Where an AI-assisted process offers little or no meaningful improvement, its ethical, environmental, material, financial, privacy, or cognitive costs may outweigh its benefits. In such cases, choosing not to use AI is not a failure to innovate but a rational and responsible decision.

The journal does not treat speed, productivity, or cost reduction as sufficient reasons for adopting AI. Efficiency is one consideration among many, not the overriding measure of scholarly or educational value.

The Value of Human Practice

Efficiency is not the only measure of academic value. Some activities may be performed effectively, or even more rapidly, with AI while still retaining greater intellectual, educational, relational, or professional value when undertaken by people. Close reading, interpretation, sustained writing, translation, feedback, mentoring, scholarly discussion, and other forms of intellectual labour can be valuable partly because of the human processes through which they are carried out.

No Presumption in Favour of Automation

The possibility of automating an academic task does not create an obligation to automate it. A task may be technically suitable for AI assistance while still being better performed by a person because the process itself carries educational, intellectual, relational, professional, or cultural value. Decisions about automation should therefore consider not only whether AI can perform a task effectively, but also what may be lost when human participation in that task is reduced or removed.

Good Faith and Equal Standards

The journal approaches both AI use and AI non-use in good faith. The use of artificial intelligence is not in itself evidence of academic misconduct, just as the decision not to use it is not evidence of methodological superiority. Submissions are assessed according to the same scholarly standards of originality, accuracy, argumentation, evidence, transparency, and accountability regardless of the tools used in their preparation. Neither responsible AI use nor principled non-use will advantage or disadvantage an author in editorial evaluation.

Responsible Criticism, Evidence & Due Process

AI & Antiquity strongly rejects the use of suspected or alleged AI use as a means of harassment, intimidation, insinuation, public shaming, or reputational harm. The mere suspicion that a text, image, argument, or other scholarly output may have involved artificial intelligence does not establish misconduct.

Legitimate concerns about undisclosed AI use, fabrication, plagiarism, authorship, manipulation of evidence, or other breaches of research integrity should be supported by relevant evidence and communicated through the journal’s established editorial procedures or, where appropriate, the corresponding institutional channels. Concerns will be evaluated on the substance of the alleged breach rather than on attitudes towards artificial intelligence itself.

No adverse editorial conclusion should be drawn solely from writing style, vocabulary, perceived fluency, stylistic regularity, personal intuition, or the output of an automated AI-detection system. Where a substantive integrity concern arises, the person concerned should normally be given an appropriate opportunity to respond before conclusions are reached.

Scholarly criticism of artificial intelligence — including strong criticism of its ethical, environmental, economic, epistemological, political, or educational implications — is fully legitimate. What the journal rejects is the substitution of unsupported personal accusation for evidence-based criticism and established procedures of academic integrity.

AI in Ancient Studies

 

Ancient Studies provides an especially valuable environment in which to examine both the possibilities and the limitations of artificial intelligence. Historical knowledge depends on incomplete evidence, contested interpretations, linguistic expertise, contextual reasoning, and careful distinctions between what sources preserve and what scholars reconstruct. These conditions make the uncritical acceptance of generated answers particularly problematic, but they also make AI a productive object through which to teach historical method and critical literacy.

Research

  • Exploratory analysis of textual, archaeological, or historical datasets.
  • Assistance with coding, classification, transcription, or information organisation.
  • Critical comparison between generated interpretations and primary or secondary sources.
  • Study of bias, historical representation, and knowledge production in generative systems.

Teaching

  • Adaptive explanations and structured learning pathways.
  • Multimodal resources, summaries, glossaries, and language support.
  • Activities that require students to identify errors, assumptions, and fabricated evidence.
  • Inclusive teaching practices that use AI to improve accessibility without lowering academic expectations.

Pedagogy Before Technology

 

The presence of a new technology does not by itself justify its educational use, and greater efficiency does not automatically constitute greater educational value. At AI & Antiquity, technological innovation is meaningful only when it addresses a genuine scholarly or pedagogical need and when its benefits are proportionate to its ethical, material, environmental, and intellectual costs. We therefore encourage authors to move beyond demonstrations of what a tool can do and to ask what it actually contributes to learning, historical understanding, accessibility, engagement, or critical thinking.

This principle is particularly important in discussions of inclusion. AI may assist with multimodal presentation, vocabulary support, flexible learning pathways, planning, or the reduction of unnecessary cognitive load. Such uses should enhance students’ capacity to engage critically with Ancient Studies rather than replace demanding intellectual work with automated answers. There will also be contexts in which the educational value lies precisely in performing a task without automation.

Limits and Risks

 

Generative AI systems can produce fluent and persuasive outputs while remaining factually unreliable. In Ancient Studies, this may include invented references, inaccurate translations, conflation of sources, chronological errors, false certainty, or the reproduction of outdated interpretations. Their apparent confidence can therefore obscure the distinction between evidence, inference, and fabrication.

  • AI-generated citations and bibliographical references must never be assumed to be genuine.
  • Translations of ancient languages require appropriate linguistic and philological verification.
  • Generated historical reconstructions should distinguish evidence from conjecture.
  • Sensitive, confidential, personal, or unpublished research material should not be uploaded to systems that cannot guarantee appropriate data protection and confidentiality.
  • Bias, unequal access, commercial dependency, copyright, labour conditions, and environmental impact are legitimate considerations when evaluating the desirability of AI-assisted practices.
  • Efficiency should not be confused with scholarly quality: faster production is valuable only when accuracy, intellectual integrity, and the purposes of the activity are preserved.

Authorship, Disclosure & Verification

 

AI systems cannot be listed as authors or co-authors of work published in AI & Antiquity. Authorship entails the capacity to defend arguments, respond to criticism, accept responsibility for errors, declare conflicts of interest, and be accountable for the integrity of the scholarly record.

Authors remain fully responsible for all submitted material, irrespective of whether artificial intelligence assisted its preparation. Substantial AI use must therefore be disclosed during submission, including, where relevant, the system used, its purpose, and the parts of the research or writing process in which it played a significant role.

Disclosure is not understood as an admission of improper practice. On the contrary, transparent reporting allows readers and reviewers to evaluate digital methods appropriately and contributes to the development of more rigorous standards for AI-assisted scholarship.

Editorial and Peer-Review Practice

 

Editorial decisions at AI & Antiquity remain human decisions. Automated systems may assist particular technical or administrative tasks, but they do not determine whether a manuscript is accepted, rejected, or revised.

The journal does not regard automated AI-detection software as sufficient evidence of undisclosed AI use or academic misconduct. No adverse editorial conclusion should be based solely on an AI-detector score or on subjective impressions that a text appears to have been generated by artificial intelligence. Concerns regarding originality or research integrity must instead be assessed through relevant evidence, editorial judgement, appropriate screening procedures, and, where necessary, direct communication with the person concerned.

Where a substantive concern arises, the journal will distinguish between the use of AI itself and any specific alleged breach of scholarly standards. It is fabrication, plagiarism, misrepresentation, failure to disclose when disclosure is required, manipulation of evidence, or another identifiable breach of research integrity that must be evaluated — not the mere presence or suspected presence of artificial intelligence.

Unpublished manuscripts, reviewer reports, editorial correspondence, personal data, and other confidential materials must not be entered into external AI systems where doing so would compromise confidentiality, intellectual property, privacy, or data protection. Editors and reviewers remain personally responsible for preserving the confidentiality entrusted to them, irrespective of the tools they use.

Technological Neutrality

 

The journal’s openness to research and experimentation involving artificial intelligence does not constitute endorsement of any particular company, commercial platform, model, technical architecture, or vision of technological development. AI systems should be evaluated according to their scholarly usefulness, limitations, accessibility, governance, ethical implications, and effects in specific contexts rather than according to novelty, market prominence, or promotional claims.

Likewise, the presence of AI in the journal’s title should not be interpreted as an assumption that artificial intelligence is intrinsically beneficial, inevitable, or desirable in every domain of Ancient Studies. It identifies an important contemporary field of inquiry whose possibilities and consequences merit sustained critical examination.

A Critical Research Agenda

 

AI & Antiquity particularly welcomes research that moves beyond simple demonstrations of generative tools and examines their broader implications for Ancient Studies. We are interested in work that tests AI critically, situates technological experimentation within disciplinary methods, evaluates both successful and unsuccessful applications, and questions assumptions about technological progress where appropriate.

Historical Knowledge

How do AI systems represent ancient societies, events, identities, historical uncertainty, and contested interpretations?

Sources & Languages

What happens when generative systems interpret ancient texts, inscriptions, archaeological evidence, material culture, or historical languages?

Education

When does AI strengthen learning, accessibility, autonomy, and critical thinking, and when does it weaken or displace valuable human practices?

Ethics & Society

How do bias, commercial infrastructures, copyright, environmental cost, labour, privacy, unequal access, and governance shape AI-assisted scholarship?

Our Guiding Principle
 

For AI & Antiquity, responsible innovation begins with the recognition that technology does not determine academic values. Historical method, critical engagement with evidence, disciplinary expertise, ethical reflection, and human judgement remain the standards against which artificial intelligence must be assessed. Our aim is not to normalise AI uncritically, nor to reject it reflexively, but to create a scholarly space in which its possibilities, limitations, and consequences can be examined openly and rigorously.

That commitment includes respect for those who choose not to use AI, resistance to the weaponisation of unsubstantiated allegations about its use, and recognition that some forms of human intellectual work remain valuable precisely because they are human. Responsible innovation requires knowing not only how to use a technology, but also when not to use it.

The journal does not require consensus on artificial intelligence, nor does participation in its scholarly community imply endorsement of AI. It requires that disagreement about it remain scholarly: open to evidence, criticism, revision, and genuine intellectual pluralism.

This statement reflects the journal’s current position in an area of rapid technological and scholarly change. It will be reviewed periodically in response to developments in research practice, publishing ethics, regulation, technology, and the continuing debate within our academic community.