AI Use & Disclosure Pack
Five copy-and-paste templates for disclosing AI use in journalism honestly and consistently - an article AI-use note, a newsroom AI-tools policy, a reader explainer, a synthetic-media label, and an AI-use log entry.
Last reviewed: Next review due:
Who this pack is for
This pack is for reporters, editors, and newsroom managers who use AI tools in their workflow and want to be transparent about it without over-claiming or under-disclosing. As transcription, translation, summarising, and generative tools become routine, readers increasingly expect to know where they were involved - and regulators expect the same standard of accuracy whether or not a tool was used.
The templates are grounded in the accuracy and transparency expectations of the IPSO Editors’ Code and, for broadcast, the Ofcom Broadcasting Code. They cover the five documents most newsrooms need: a note for individual articles, a policy for the newsroom, a standing explainer for readers, a label for synthetic media, and an internal log entry so you can show what you did. Adapt them to your regulator, your tools, and your house style.
What’s in this pack
Five templates - copy, adapt, and use.
Template 1: Article AI-Use Note
A short, reader-facing note for the foot of a story (or an editor note box) that states plainly how AI helped and confirms a journalist checked the facts.
Template 2: AI-Tools Editorial Policy Statement
A newsroom policy that sets out permitted uses, sign-off requirements, prohibited uses, and verification duties, anchored to your regulator accuracy standards.
Template 3: Reader-Facing "How We Use AI" Explainer
A plain-English standing page telling readers where AI does and does not feature in your journalism, and how to raise a correction.
Template 4: AI Image / Synthetic-Media Disclosure Label
Caption and credit-line wording for AI-generated, AI-altered, or AI-restored images and audio. Pick the variant that matches what was actually done.
Template 5: Newsroom AI-Use Log Entry
An internal record of a single instance of AI use, capturing the tool, task, inputs, verification, and whether it was disclosed - your evidence trail if a decision is later questioned.
Template 1: Article AI-Use Note
A short, reader-facing note for the foot of a story (or an editor note box) that states plainly how AI helped and confirms a journalist checked the facts.
AI-USE NOTE - ARTICLE FOOTER [Append to the foot of the article, or place in an editor's note box near the top.] AI-use note: This article was researched and written by [JOURNALIST NAME]. [Select and adapt the lines that apply; delete the rest.] - We used [AI TOOL / TYPE, e.g. a large language model] to [SPECIFIC TASK, e.g. transcribe a recorded interview / summarise a public dataset / suggest subheadings]. - All facts, quotes, and figures were checked by a journalist against primary sources before publication. - No AI system was used to generate quotes, sources, or evidence presented as fact. - [If applicable] An image or audio clip in this article was generated or altered using AI; see the label on that item. Editorial responsibility for this article rests with [NAME / TITLE] at [PUBLICATION]. If you believe anything here is inaccurate, contact [CORRECTIONS EMAIL] and we will review it promptly.
Template 2: AI-Tools Editorial Policy Statement
A newsroom policy that sets out permitted uses, sign-off requirements, prohibited uses, and verification duties, anchored to your regulator accuracy standards.
AI TOOLS - EDITORIAL POLICY STATEMENT [PUBLICATION] policy on the use of artificial intelligence in journalism Version: [X] | Effective date: [DATE] | Owner: [EDITOR / TITLE] 1. Principle. We use AI tools to support journalists, never to replace editorial judgement. Accuracy, fairness, and transparency under the IPSO Editors' Code [or: the [REGULATOR] code] apply to AI-assisted work exactly as they do to all our journalism. 2. Permitted uses. Journalists may use approved tools for: transcription, draft translation, summarising documents we already hold, generating research leads (to be independently verified), copy-editing suggestions, and headline or structure options. 3. Uses that require sign-off. The following need prior approval from [EDITOR / TITLE]: publishing any AI-generated or AI-altered image or audio; using AI to draft substantive article text; and any use involving personal data or confidential source material. 4. Prohibited uses. We do not present AI-generated text, quotes, or sources as verified fact without checking; we do not enter confidential source information or unpublished investigation material into third-party tools; and we do not use AI to fabricate or "enhance" evidence, images, or events. 5. Verification. A journalist is responsible for every fact in their copy. AI output is a draft or a lead, never a source. All names, figures, quotes, and claims must be checked against primary sources. 6. Disclosure. Where AI use is material to how a story was produced, we disclose it to readers. Synthetic or materially altered images and audio are always labelled. 7. Record-keeping. Significant AI use is recorded in the newsroom AI-use log (tool, task, journalist, date, verification status). 8. Review. This policy is reviewed at least annually, and sooner when new tools or regulatory guidance require it. Approved by: [NAME / TITLE] Date: [DATE]
Template 3: Reader-Facing "How We Use AI" Explainer
A plain-English standing page telling readers where AI does and does not feature in your journalism, and how to raise a correction.
HOW WE USE AI - READER EXPLAINER [Publish on a standing "About our journalism" or "Editorial standards" page.] How [PUBLICATION] uses artificial intelligence We want to be open about where AI does, and does not, play a part in our journalism. What AI helps us with: - Transcribing interviews and drafting translations of source material, which a journalist then checks. - Searching and summarising large documents and datasets we are investigating. - Suggesting headlines, subheadings, and structure, which our editors rewrite and approve. What AI does not do: - It does not decide what we report or write our stories for us. A named journalist is responsible for every article. - It does not create quotes, sources, or evidence. We verify facts against primary sources. - We do not publish AI-generated or AI-altered images or audio without clearly labelling them. Our standards. Our journalism follows the IPSO Editors' Code [or: the [REGULATOR] code] on accuracy and fairness, whether or not a tool helped produce it. Corrections. If we get something wrong, we correct it promptly and transparently. Email [CONTACT EMAIL]. Last updated: [DATE]
Template 4: AI Image / Synthetic-Media Disclosure Label
Caption and credit-line wording for AI-generated, AI-altered, or AI-restored images and audio. Pick the variant that matches what was actually done.
SYNTHETIC / AI MEDIA - DISCLOSURE LABEL [Place in the caption, credit line, or an on-image overlay. Choose the wording that matches what was done and delete the rest.] FULLY AI-GENERATED IMAGE Illustration generated using AI [TOOL, if disclosed]. This image does not depict a real person, place, or event. Created for [PUBLICATION] by [NAME]. AI-ALTERED PHOTOGRAPH This photograph has been altered using AI [describe change, e.g. background removed / elements added]. Original photograph: [PHOTOGRAPHER / AGENCY]. AI-GENERATED OR CLONED VOICE The voice in this clip was generated using AI and is not a recording of the named person speaking these words. [Explain, e.g. used to voice a written statement with the subject's consent]. AI-RESTORED OR UPSCALED ARCHIVE MATERIAL This archive image has been AI-restored or upscaled. Detail not present in the original may have been reconstructed by the tool. Rule: never use an AI-generated image to depict a real, identifiable news event as though it were a genuine photograph of that event.
Template 5: Newsroom AI-Use Log Entry
An internal record of a single instance of AI use, capturing the tool, task, inputs, verification, and whether it was disclosed - your evidence trail if a decision is later questioned.
NEWSROOM AI-USE LOG - ENTRY Log ID: [SEQUENTIAL NUMBER] Date: [DATE] Journalist / desk: [NAME / DESK] Story or project: [SLUG OR HEADLINE] Tool used: [TOOL NAME AND VERSION / TYPE] Task performed: [e.g. transcription of a 40-minute interview / summary of a 200-page report / three headline options] Inputs: [What was given to the tool. Confirm no confidential source data or unpublished investigation material was entered - YES / NO. If NO, explain and escalate to the editor.] Output used how: [e.g. transcript checked against audio and used for quotes / summary used as a research lead only / one headline option edited and used] Verification: [Who checked the output, against what primary source, and when.] Verified by: [NAME] Date: [DATE] Disclosed to readers? [YES - via AI-use note / image label] [NO - not material] Sign-off (if required): [EDITOR NAME] Date: [DATE]
Primary sources
- IPSO Editors' Code of Practice - Clause 1 (Accuracy) applies to AI-assisted journalism exactly as to any other reporting
- NUJ Code of Conduct - accuracy, honesty about how information was obtained, and not distorting material
- Ofcom Broadcasting Code - accuracy and fairness rules that govern AI-generated or altered content in broadcast
- ICO - Artificial Intelligence guidance - data-protection considerations before putting personal or confidential data into AI tools