Skip to main content

SPIE Policy on Artificial Intelligence and Generative AI

In This Section

Disclosure and Responsibility

SPIE recognizes that artificial intelligence (AI), including large language models (LLMs), is increasingly used throughout the research and writing process. AI tools may be used in ways that support efficiency and clarity without compromising scholarly standards. This policy focuses on accountability, reproducibility, and integrity, rather than exhaustive technological disclosure.

During journal submission, authors complete a structured AI‑use declaration by indicating whether AI or automated tools were used in any of the following categories:

  • Code generation or debugging
  • Data processing or statistical analysis
  • Figure or image generation or modification
  • Language editing, translation, or grammar support
  • Literature discovery or summarization
  • Modeling or simulation

Authors are not required to:

  • Describe detailed workflows
  • Provide prompts, prompt logs, or conversational transcripts
  • Justify routine AI use for writing or editing

Uses of AI

SPIE recognizes that AI and automated tools may be used at multiple stages of the research and manuscript preparation process. The relevance of such use to editorial review depends on whether it is integral to the methods or results being reported.

AI use relevant to methods and results

When AI or automated tools are used in activities that contribute directly to the generation, analysis, or presentation of research findings, such as:

  • data collection or preprocessing,
  • statistical analysis or inference,
  • modeling, simulation, or prediction,
  • code generation that supports results,
  • figure or image construction tied to findings,

editors may request additional methodological clarification, consistent with existing expectations for computational, statistical, or analytical methods.

Such clarification may include, where relevant:

  • identification of the tool or system used
  • description of how outputs were reviewed, validated, or quality checked
  • explanation sufficient for an expert reader to understand the role of the tool in the study and to assess the reliability and reproducibility of the results.

These requests are intended to support transparency and reproducibility and do not imply that the use of AI tools is improper or disfavored.

AI use for writing and presentation support

AI or automated tools used solely for assistance with:

  • language editing,
  • grammar or clarity improvement,
  • translation,
  • stylistic refinement,

do not normally require further editorial review beyond the initial disclosure made during submission. Such use does not usually warrant description in the Methods section and need only be mentioned elsewhere if authors consider it substantively relevant.

Author Responsibility and Accountability Statement

All corresponding authors must affirm the following prior to publication:

  • I take full responsibility for the data, analyses, results, and interpretations reported in this manuscript.
  • I confirm that this work complies with SPIE's standards for reproducibility, transparency, and research integrity as described in the SPIE Guidelines for Ethical Publishing.

Prohibited Uses

AI tools must not be used in ways that:

  • Fabricate or falsify data, results, references, or citations
  • Disguise unsupported data or lack of methodological rigor
  • Obscure authorship responsibility
  • Violate confidentiality in peer review (see Reviewer Guidelines).

Undisclosed AI use that materially affects results or interpretation may be treated as a research integrity concern.

Recommended Best Practices for Authors

SPIE encourages the following practices:

  • Verification: Independently check AI-assisted outputs, especially code, statistics, and references. False outputs fabricated by AI are cause for rejection.
  • Version control: Retain records of analytical steps, data transformations, and code versions, consistent with standard computational workflows.
  • Transparency by design: When AI meaningfully shapes results, document its role as you would any other method, at a level sufficient for expert evaluation.
  • Human oversight: Use AI to aid, not replace, author expertise and scientific judgment.

Recommended Best Practices for Reviewers

Reviewers are expected to follow the same ethical standards regardless of whether AI tools are used for limited assistance.

  • Maintain strict confidentiality. Manuscripts, figures, data, and review comments must not be entered into web-based LLMs. Doing so constitutes a breach of peer-review confidentiality.
  • Do not delegate the review to AI tools. AI tools may be used to assist with review preparation, including drafting or evaluation, provided that reviewers independently assess, verify, and agree with all content. Responsibility for the substance and recommendations of the review rests entirely with the reviewer.
  • Permitted, limited AI uses. Reviewers may use AI tools for:
    • High-level background summaries of a field
    • Language or stylistic polishing of their own review text
    provided no manuscript content is uploaded and confidentiality is preserved.
  • Validate all AI-assisted output. Any information obtained via AI tools must be independently verified. Reviewers are responsible for accuracy, relevance, and appropriate use in their reviews.
  • Disclose AI use to the editor. Any use of AI tools in the review process must be disclosed in the "Private Comments to Editor" field, as required by Reviewer Guidelines.
  • Raise content concerns, not AI suspicions. If results seem unclear, references incorrect, or methods incomplete, reviewers should flag these issues without speculating about AI involvement.

Recommended Best Practices for Editors

Editors are responsible for overseeing peer review in a manner that safeguards confidentiality, fairness, and research integrity, including when AI or automated tools are used.

  • Apply AI policies consistently with existing reviewer obligations. Editorial expectations regarding AI use should align with the Reviewer Guidelines, particularly confidentiality and disclosure requirements. Editors should not introduce additional requirements related to AI use beyond established policy language.
  • Focus on integrity risks, not stylistic signals. Editorial attention should center on verifiable concerns such as insufficient methods, irreproducible analyses, or inaccurate references. Writing style or perceived AI use alone is not evidence of a policy violation.
  • Enforce confidentiality strictly in peer review. If there is reason to believe a reviewer uploaded a manuscript to a public, web-based LLM, editors should treat the issue as a confidentiality breach, consistent with Reviewer Guidelines.
  • Use disclosure, not detection, as the primary control. Editors should rely on required reviewer disclosures of AI use rather than attempting to infer AI involvement indirectly.
  • Assess reviewer AI use by outcome and compliance. The key questions for editors are whether the review was independently reasoned, confidential, and scientifically sound—not which tools assisted background understanding or language refinement.
  • Escalate only when policy thresholds are crossed. Allowable AI use in review (e.g., summarizing background literature) should not trigger action. Prohibited use (e.g., uploading manuscripts or review content to external AI systems) should be addressed using existing reviewer conduct procedures.