Artificial Intelligence Use Policy
Penn State University Press
I. Purpose and Institutional Commitments
Penn State University Press (PSUP) is a public, land-grant, scholarly publisher committed to advancing reliable knowledge, intellectual independence, and public trust in the scholarly record. As a publisher specializing in the humanities and social sciences, PSUP recognizes that its core mission is to preserve, disseminate, and advance works grounded in human interpretation, critical judgment, moral reasoning, and cultural understanding. These disciplines, along with the creative works published under PSUP imprints—including prose fiction and graphic narrative—depend fundamentally on the expression of the human intellect, which remains the primary engine of both scholarly discovery and creative authorship.
Clear and enforceable policies governing AI Technology are therefore necessary to ensure that scholarly and creative work published by PSUP reflects genuine human inquiry rather than machine output. While acknowledging the potential for these tools to support research efficiency and creativity, the PSUP identifies significant risks associated with current generative systems, including inaccuracies, algorithmic bias, false authenticity, and the erosion of precise attribution standards. This policy governs the responsible use of computational tools by all entities engaged in work for PSUP to preserve intellectual agency and protect the integrity of the scholarly record, and guarantee the originality of expression. AI Technology may support scholarly and creative work published at PSUP; however, it may not replace human credit, responsibility, judgment, or authorship.
II. Definitions
To ensure clarity and consistency across all PSUP operations, the following terms are formally defined:
1. AI Technology: Any generative artificial intelligence, large language model (LLM), or multimodal computational system capable of producing, synthesizing, or substantially transforming text, data, or imagery through machine learning. This definition does not encompass routine productivity tools whose primary function is mechanical or rule-based assistance, including spell checkers, standard grammar correction software, optical character recognition (OCR) utilities, and basic formatting aids. Such tools do not require Disclosure under this policy.
2. Covered Materials: All manuscripts (submitted or unsubmitted, finished or unfinished, published or unpublished), editorial communications, peer review reports, research data, and visual materials associated with PSUP publications or internal processes.
3. Covered Parties: All individuals and entities engaged in work for PSUP, including authors, contributors, editors, reviewers, editorial board members, contractors, vendors, and employees.
4. Substantive AI Contribution: The use of AI Technology to generate, substantially draft, or structurally reorganize textual content within a work prepared for or submitted to PSUP, including prose, argumentative framing, narrative passages, literature summaries, and other written elements. The term also includes the use of AI Technology to analyze or translate research data or source materials on which such a work materially relies. This term does not apply to grammar correction, spell-checking, or light copyediting, even when performed with AI Technology, and such uses do not require Disclosure. A Substantive AI Contribution requires Disclosure under Section VI.1.
5. Visual Enhancement: The use of AI Technology to substantively introduce, remove, or alter content within visual features—such as charts, graphs, images, photos, and archival scans—through generative fill, synthetic reconstruction, or compositing. This includes any synthetic transformation that changes the narrative context or evidentiary value of the visual record. It excludes routine, non-generative adjustments like cropping, resizing, or standard color balancing.
6. Disclosure: A formal, good-faith declaration to PSUP identifying the specific tool(s) used, its functional role in the workflow, and the nature of human oversight applied to the output.
III. Scope and Applicability
This policy applies to all manuscripts, editorial processes, reviews, services, and systems associated with PSUP publications. Compliance with these standards is a mandatory condition for manuscript submission, publication consideration, peer review participation, editorial activity, and contractual engagement. All Covered Parties are expected to adhere to these standards in any work performed for or on behalf of PSUP. For journals not owned by PSUP, this policy applies except where the owning organization has adopted its own AI policy (see Section VIII.3). This policy takes effect on January 1, 2027, and applies to any project contracted with PSUP on or after that date.
IV. Core Principles
1. Human Intellectual Primacy
All scholarship and creative work—especially within the humanities and social sciences—is inseparable from human interpretation, ethical reflection, and intellectual risk-taking. PSUP affirms that the expression of the human intellect is not merely a stylistic preference but a defining feature of legitimate scholarly work. Because AI Technology lacks moral agency, scholarly judgment, and emotional intelligence, it may assist cognition but may not substitute for intellectual authorship or interpretive responsibility.
2. Transparency and Disclosure
Any Substantive AI Contribution or Visual Enhancement must be disclosed accurately, consistently, and in good faith. Covered Parties must provide Disclosure of uses that materially inform the generation, synthesis, or reorganization of any work prepared for, submitted to, or published by PSUP, including any Visual Enhancement of Covered Materials.
3. Accountability and Liability
Authors and editors remain fully responsible for the accuracy, originality, and ethical compliance of their work. Responsibility for truth and error cannot be delegated to automated systems. Covered Parties are accountable for reviewing all machine-generated outputs to confirm content accuracy and the validity of citations.
4. Rights Protection and Data Privacy
The use of AI Technology must align with privacy, confidentiality, and intellectual property obligations. Covered Parties must ensure that the tools they select do not claim ownership of Covered Materials or impose limitations that could interfere with PSUP’s rights to publish and distribute the work.
V. Prohibited Delegations (“Red Lines”)
The following functions are essential to human authorship and may never be delegated to AI Technology by authors or editors. For purposes of this section, “delegation” means accepting AI-generated output as a substitute for one’s own scholarly reasoning—whether directly or after only superficial review. Using AI as a generative aid in a process that culminates in substantive, independent human judgment does not constitute delegation; presenting AI output as the product of that judgment does.
·Original Scholarly Judgment: This includes the interpretation of archival sources, weighing competing theories, and causal analysis.
·Formulation of Research Questions: Framing scholarly problems and defining the scope of inquiry must remain a human endeavor.
·Ethical Evaluation: Moral claims, justice analysis, and cultural critiques require human empathy and social context.
·Responsibility for Truth: Verifying sources, checking data for errors, and correcting factual mistakes is a human mandate.
·Engagement with Primary Sources: The outsourcing of archival reading, ethnographic interpretation, or close textual analysis is prohibited.
·Original Creative Authorship: The generation of narrative prose, fictional voice, character development, or plot construction must remain the product of human creative agency. AI may assist with drafting or revision, subject to Section VI, but the animating creative vision of a work must be the author’s own.
·Visual and Sequential Storytelling: For illustrated and graphic works, the artistic decisions that constitute the work—including character design, panel composition, visual narrative structure, and illustrative style—may not be delegated to AI Technology. The visual authorship of a work carries the same protections as its written authorship.
·Translation of Press Content: The wholesale translation of any PSUP publication or manuscript into another language may not be delegated to AI Technology. Translation of scholarly and creative works requires human interpretive judgment, cultural knowledge, and fidelity to authorial voice that automated systems cannot reliably supply. AI Technology-assisted translation tools may support a human translator’s process but may not substitute for one.
Delegation of these core responsibilities constitutes a material breach of PSUP policy and undermines the validity of the scholarly record in the humanities and social sciences.
VI. Standards for Authors and Contributors
1. Mandatory Disclosure Protocols
All submissions must include a formal Disclosure of any Substantive AI Contribution or Visual Enhancement, or an explicit declaration that neither applies. Routine, non-generative tools—such as spell checkers, grammar correction software, and OCR utilities—fall outside the definition of AI Technology and do not require Disclosure. Authors are encouraged to use the GAIDeT Declaration Generator ** or similar formal instruments to standardize these declarations to PSUP.
Disclosure does not automatically count against a submission. PSUP evaluates a disclosed Substantive AI Contribution on its merits, as it evaluates every other element of a submission, with attention to the nature and extent of the use and its implications for the copyrightability of the work. Disclosure does not, however, permit a delegation prohibited under Section V.
A Substantive AI Contribution must also be acknowledged in the published work, with a brief explanation: in a book's preface, introduction, or acknowledgements, or in a journal article’s acknowledgements. Visual Enhancements are disclosed as provided in Section IX.1.
Material misrepresentation regarding the use of AI Technology constitutes grounds for manuscript rejection, contract termination, or retraction, as PSUP deems appropriate.
2. Authorship and Legal Standing
AI Technology cannot fulfill the requirements for authorship because it cannot assume responsibility for content, consent to publication agreements, or manage copyright licenses. Consequently, such tools must never be listed as an author or co-author on any PSUP publication. Authorship requires uniquely human responsibilities that cannot be undertaken by automated systems.
3. Verification of Outputs and References
Authors must carefully verify the impartial nature and accuracy of all machine-assisted outputs. Because AI Technology may generate incorrect or fabricated references, authors must manually check every citation against original sources. The final work must reflect the author’s authentic expertise, voice, and original interpretation.
4. Recordkeeping Requirements
Authors are expected to maintain documentation of the models used, prompt categories employed, and general workflow structures for any use requiring Disclosure.
At minimum, documentation must include:
·The name and version of each AI Technology tool used
·The general category of task for which it was used (e.g., drafting, literature summarization, visual generation)
·A brief description of the nature of human oversight applied to each output
Retention period: Documentation must be retained at least through the date of formal submission to PSUP and, if the work is accepted, through the date of publication. Authors are strongly encouraged to retain records indefinitely. The scholarly record is subject to scrutiny at any point after publication, and the ability to reconstruct one’s workflow may be essential to defending the integrity of the work against future allegations or corrections requests. Documentation that cannot be produced upon request during the required retention period may itself be treated as evidence of non-compliance.
5. Use of This Document as a Reference Aid
Authors are expressly permitted to upload this policy (this document) to AI Technology for the purpose of asking questions about what uses are allowable, permissible, or required under PSUP policy. Authors are encouraged to use this method as a self-service compliance aid before, during, and after the preparation of their submissions. This document is designed to be machine-readable for that purpose.
VII. Standards for Reviewers
1. Absolute Confidentiality
Reviewers are strictly prohibited from uploading any portion of Covered Materials under review—including manuscripts, figures, or tables—to AI Technology for any purpose. This prohibition is absolute and serves to protect the author’s intellectual property and the integrity of the blind review process.
2. Evaluation Integrity
Reviewers must be human and must engage directly with submitted texts to derive independent, human-created assessments. PSUP does not permit the use of AI Technology to analyze, summarize, or generate the evaluative portions of a review. The critical thinking required for peer review is currently outside the scope of reliable machine processing.
3. Permissible Support
Reviewers may use AI Technology to assist with improving the language and readability of their final reports. However, this use must be transparently declared to the handling editor upon submission. Reviewers remain fully responsible for the accuracy and integrity of their assessments regardless of tool usage.
VIII. Standards for Editors
1. Preservation of Human Judgment
Editors may not use AI Technology to rank submissions, predict the impact of a work, or recommend acceptance or rejection. All editorial determinations must be made by human reasoning and critical assessment. Delegating the evaluation of a scientific or humanistic manuscript to an automated system poses risks of generating incomplete or biased conclusions.
2. Confidentiality in Communication
Editors must treat all submitted manuscripts as confidential documents. This confidentiality requirement extends to decision letters and notifications, which often contain sensitive information about the manuscript or its authors. Editors should not upload these communications to AI Technology, even for language polishing, except as authorized for PSUP employees under Section X.2.
3. Policy Oversight and Reporting
Editors are expected to enforce PSUP’s AI policies within their respective publications. If an editor suspects that an author or reviewer has violated these standards, they are obligated to inform PSUP leadership for further investigation. For journals not owned by PSUP, governance resides with the owning organization.
IX. Visual Materials and Artwork
1. Disclosure for Enhanced Materials
For creative and graphic works, visual elements do not serve merely an evidentiary function—they may constitute the work itself. AI-generated imagery, character likenesses, panel compositions, or illustrative content used in such works must be disclosed with the same specificity required of any other Visual Enhancement. PSUP retains editorial discretion to accept or decline any submission containing such content based on the nature and extent of its use. Any image, graphic, or visual element that has been generated or subjected to Visual Enhancement must include a Disclosure identifying the specific platform and model version used. This requirement applies to all contributors, including authors, designers, and third-party contractors. Disclosures must appear in the image credit line or a location designated by PSUP.
2. Prohibited Manipulations
PSUP does not permit the use of AI Technology to create or alter original research data, images, or figures in ways that move, remove, or conceal specific features. Creating likenesses of real people or identifiable brands without authorization is strictly prohibited.
3. Forensics and Exceptions
PSUP may utilize image forensic software or specialized tools to identify suspected irregularities in Covered Materials. Routine, non-generative adjustments—such as cropping, resizing, or standard color balancing—do not require Disclosure.
X. Vendor and PSUP Employee Standards
1. Restrictions for Vendors
Absent explicit written permission from PSUP, vendors, freelancers, suppliers, and platform partners are prohibited from using AI Technology on Covered Materials. Vendors are expressly barred from uploading Covered Materials to AI Technology of any kind, be it public-facing, private, or a “walled garden” environment. Vendors may not distribute or reuse Covered Materials beyond the agreed-upon scope of work.
2. Employee Internal Operations
PSUP employees are subject to internal policies designed to protect the confidentiality of Covered Materials entrusted to PSUP. As a general principle, employees may not provide Covered Materials to external AI Technology systems unless such use has been explicitly authorized by the Executive Director. Permitted uses will only include environments that maintain Penn State University institutional control over data and prevent its incorporation into third-party training sets.
XI. Enforcement and Review
1. Consequences of Non-Compliance
Material violations of this policy may result in the rejection of manuscripts, retraction of published work, termination of agreements, or public correction of the scholarly record. In severe cases, PSUP may initiate institutional reporting or disassociation. The severity of enforcement shall be determined by PSUP based on the scope, intent, and impact of the violation. Before taking any such action other than declining a submission, PSUP will notify the affected party of the suspected violation and provide a reasonable opportunity to respond.
2. Authority and Review Cycle
Interpretation and enforcement authority for this policy resides with the Executive Director of PSUP. To remain responsive to the rapid evolution of AI Technology, this policy shall be reviewed and revised at least every three years.
** Suchikova, Y., Tsybuliak, N., Teixeira da Silva, J. A., & Nazarovets, S. (2025). GAIDeT (Generative AI Delegation Taxonomy): A taxonomy for humans to delegate tasks to generative artificial intelligence in scientific research and publishing. Accountability in Research, 1–27. https://doi.org/10.1080/08989621.2025.2544331
Fall/Winter 2026