What Should Researchers Do When University and Journal AI Policies Conflict?

What Happens When Two AI Policies Tell You Different Things?

Your university permits a particular form of AI-assisted writing, but your target journal restricts it. Or the journal asks authors to disclose AI use in a way that seems inconsistent with your institution's confidentiality requirements. Perhaps your university requires an approved AI platform while the journal's instructions appear to assume another workflow.

Which policy wins?

That question is tempting, but it can be misleading. University and journal policies often govern different parts of the research and publication process. Many apparent conflicts disappear once you identify precisely what each policy requires. A genuine conflict exists when two requirements both apply to the same activity and complying with one would prevent you from complying with the other.

Do Not Automatically Choose One Policy Over the Other

Most Policy Differences Are Not True Conflicts

Start by Separating Permission From Obligation

Suppose your university says researchers may use generative AI for manuscript preparation under specified conditions. Your journal says substantive AI assistance must be disclosed.

Those policies do not conflict. One grants conditional permission. The other imposes an additional publication requirement. You can use the tool in accordance with the university policy and disclose that use in accordance with the journal policy.

Researchers sometimes interpret "the university allows this" as "no other authority may impose additional conditions." That is not what permission ordinarily means.

This distinction should be made before attempting to decide which policy has priority.

Identify Exactly What Each Policy Governs

University policies can govern approved software, research data, information security, privacy, ethics, research integrity, intellectual property, employment responsibilities, and institutional resources.

Journal policies can govern manuscripts submitted for publication, AI declarations, authorship, images, methodological reporting, confidentiality during peer review, and editorial processes.

These scopes overlap, but they are not identical. Understanding which authorities govern each part of AI-assisted research is therefore the first step in resolving an apparent conflict.

Write the Two Requirements Side by Side

Policy conflicts become much easier to analyse when vague impressions are converted into specific requirements.

Reading policies this way is less glamorous than debating "who has authority," but it usually produces a better answer. Policy interpretation has always had a fondness for verbs.

A Stricter Journal Rule Often Creates No Conflict

Imagine that your university allows AI-assisted manuscript drafting while your target journal allows AI only for specified forms of assistance and prohibits AI from replacing accountable scholarly contribution.

You can normally comply with both by limiting your AI use to what the journal permits. The university did not require you to use every capability it allows.

Current publisher policies illustrate how additional conditions can operate. Elsevier permits certain AI assistance but requires qualifying use in manuscript preparation to be declared. Springer Nature uses a risk-based framework that permits some uses, requires oversight and transparency for others, and prohibits uses that replace accountable human scholarship or create unverifiable scholarly content. ICMJE likewise emphasises human accountability, transparency, and journal-specific AI policies.

The existence of these publication requirements does not inherently contradict a university policy that permits responsible AI use.

A Stricter University Rule Also Usually Remains Applicable

Reverse the example. The journal permits authors to use generative AI to improve manuscript text, but your university prohibits confidential or sensitive research material from being entered into the AI service you intend to use.

The journal's permission does not compel you to upload the material. You can comply with both by not using that service with restricted information.

This reflects the distinction discussed when considering whether a journal can restrict AI use that your university otherwise permits: permissions from different authorities are not interchangeable.

True Conflicts Require More Careful Resolution

A genuine conflict is less common but possible. For example, a journal could request information about an AI workflow that your institution believes cannot be disclosed publicly because it contains protected security information, confidential contractual details, or proprietary material.

Even then, the conflict may concern the form or level of disclosure rather than whether disclosure is possible at all. The appropriate solution might involve providing sufficient methodological transparency without exposing protected information, but researchers should not invent such compromises without checking whether both authorities accept them.

Another possibility is that a journal requires a procedure that your institution explicitly prohibits. If there is genuinely no compliant way to satisfy both, submission to that journal may not be possible unless one authority grants an exception or acceptable alternative.

Check Whether Other Authorities Constrain the Solution

A disagreement between university and journal policies may not be the only issue. The research could also be governed by law, an ethics approval, a grant condition, a data-use agreement, an employment obligation, or a collaboration contract.

Suppose the journal permits an AI analysis and the university also permits it, but the dataset licence prohibits transferring the data to third-party services. There is no university-journal conflict at all. The contractual restriction independently prevents the proposed workflow.

Similarly, funder requirements can impose their own AI conditions, while research ethics requirements may constrain AI processing involving participants.

Ask the University About the University Rule

If the institutional requirement is ambiguous, seek clarification from the office responsible for it. Depending on the issue, that may be the research office, research-integrity office, data-protection office, information-security team, ethics office, legal office, or another policy owner.

Do not ask only, "Can I follow this journal policy?" Explain the exact apparent conflict. Identify what the journal requires, what institutional provision appears to prevent it, what information or AI system is involved, and what you propose to do.

Ask the Journal About the Journal Rule

When the publication requirement itself is unclear, contact the journal through the appropriate editorial channel. Editors are better positioned to interpret their journal's requirements than a discussion thread, colleague's recollection, or AI-generated summary of the policy.

This is particularly important where the journal's published guidance does not address your specific use. The separate problem of what to do when a journal has no clear AI policy should not be solved by inventing one on its behalf.

Document the Clarification

When the issue is consequential, keep the relevant policy versions and written clarification. AI policies are changing quickly, and a future editor, collaborator, research-integrity officer, or reviewer may reasonably ask what requirements you followed.

Documentation is particularly useful when an authorised office or editor provides an interpretation not obvious from the public wording.

Changing the Journal Can Be a Legitimate Solution

Researchers sometimes treat the target journal as immovable. It is not always so.

If a journal's requirements are genuinely incompatible with obligations you cannot waive, another appropriate publication venue may be the simplest compliant solution. This should not be confused with shopping for a journal that will overlook questionable conduct. The issue is whether another legitimate venue has requirements compatible with how the research was lawfully and ethically conducted.

Is This Actually a Policy Conflict?

This kind of distinction prevents unnecessary escalation. Before asking which policy wins, first ask whether both can be followed.

Common Mistakes When AI Policies Appear to Conflict

Resolve the Requirements Before Choosing Which Rule to Follow

When policies appear inconsistent, turn the problem into a sequence of concrete questions rather than deciding immediately that one authority must outrank the other.

For collaborative projects, make the resolution visible to the research team. A documented project-level AI-use plan can help prevent one collaborator from acting on a journal rule while another is working from a different institutional assumption.

When University and Journal AI Policies Seem to Disagree

Frequently Asked Questions About Conflicting AI Policies

First Determine Whether the Policies Really Conflict

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.