AI Use Statement

Our approach

The Liver Critical Care Hub uses artificial intelligence as a supporting editorial tool. It does not replace clinical expertise, critical appraisal or human editorial judgement.

AI is not an editor, author, reviewer or source of evidence. A named human remains responsible for every item published by the Hub, including its accuracy, interpretation, references and disclosures.

The Hub does not provide a user-facing AI clinical assistant. AI is not used to give individual medical advice or make decisions about patient care.

Where AI may be used

AI tools may assist with:

• literature surveillance and initial classification of new publications;
• developing article structures, outlines and questions for appraisal;
• drafting support;
• language editing and improving readability;
• formatting, tagging and preparation of metadata;
• summarising material for subsequent human verification; and
• administrative and editorial workflows, such as identifying duplicate items or organising a publication queue.

These uses are intended to reduce repetitive work and support consistency. They do not remove the need for a human to decide whether a topic is worth covering or what the Clinical Takeaway should be.

Human control and accountability

AI does not select content for publication autonomously and cannot move an item to Published. Topic selection, clinical judgement, interpretation and final publication decisions are made by the Founder and Editor or another named human editor.

AI tools are not listed as authors. They cannot take responsibility for accuracy, integrity, originality, conflicts of interest or errors. The named human author and editor remain accountable for all published material, including material initially produced with AI assistance.

Humans are responsible for ensuring that published work is original, appropriately attributed and does not reproduce protected material without permission.

Clinical claims and references

AI-generated output can be incorrect, incomplete, outdated or biased. It can also produce plausible references that do not exist or do not support the claim being made.

For that reason:

• important clinical claims and numerical results are checked against the primary source;
• quotations are checked against the original text;
• citations suggested by AI are verified before use;
• an AI-generated citation that cannot be confirmed is not accepted;
• AI output is not cited as the primary evidence for a clinical claim; and
• the final interpretation and practice judgement are made by a human editor or author.

Where evidence is uncertain or conflicting, AI-generated certainty will not be allowed to conceal that uncertainty.

Confidentiality and patient information

Identifiable patient information is never entered into public AI systems for Hub work.

Confidential manuscripts, unpublished research, peer-review material, private correspondence or information supplied in confidence will not be uploaded to an AI system without appropriate authority and adequate assurance that confidentiality is protected. Only the minimum information necessary for an approved task should be used.

Contributors must tell the editor if material they provide is confidential, unpublished or subject to restrictions. Patient-based material must be appropriately anonymised, and any necessary consent or permission must be obtained before publication.

Disclosure of AI involvement

This statement provides a general disclosure of the ways AI may support the Hub’s standing editorial workflow.

Where AI has materially contributed to an individual publication, for example by generating or substantially shaping text, an analytical framework, an illustration or other published content, this will be disclosed with the item or image where appropriate. The disclosure should describe the nature and purpose of the use, and identify the tool where this is relevant.

Routine assistance with spelling, grammar, formatting, metadata or administrative organisation is covered by this site-wide statement and may not receive a separate item-level disclosure.

Authors and contributors are expected to disclose material AI use when submitting work. Nondisclosure does not transfer responsibility away from the named human author.

Images and visual material

Materially AI-generated or AI-edited images will be identified where appropriate, usually in the caption or accompanying disclosure.

AI will not be used to fabricate, obscure or misleadingly alter clinical data, research figures or patient images. Factual diagrams and graphical summaries remain subject to human checking, appropriate attribution and any necessary permissions.

Editorial boundaries

AI may help identify potentially relevant evidence, but it does not determine clinical importance, methodological quality or whether practice should change. It does not replace direct reading of the source paper, guideline or opportunity listing.

Nothing produced through an AI-assisted workflow is published automatically. The Hub’s editorial statuses remain human controlled, from initial idea through source checking, review, preview and publication.

Standards informing this policy

Although the Hub is not a medical journal, this approach is informed by the International Committee of Medical Journal Editors recommendations on the use of artificial intelligence. These emphasise transparency, confidentiality, appropriate attribution and ultimate human accountability, and state that AI tools should not be listed as authors.

Questions and concerns

This statement will be reviewed as the Hub’s workflow, available technologies and relevant professional guidance evolve.

Questions or concerns about AI use can be sent to livercriticalcarehub@gmail.com. Please identify the relevant page, article or image so that the concern can be reviewed.