iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Disclose AI assistance when it materially shapes published content or when the publisher’s policy requires it. Tell readers what the AI did, what part of the work it affected, and what a person actually reviewed or verified. There is no single disclosure threshold or required wording across newsrooms, so check the rules that apply to your publication and the kind of AI use involved.
Decide whether the AI use needs a content-level disclosure
Use these questions before publication to determine whether readers need a note attached to the work:
- Did AI-generated output make it into the published text, image, audio, translation, analysis, or recommendation?
- How substantial was its contribution? Could it change the meaning, accuracy, framing, or an editorial decision?
- Would a reader reasonably consider the use material, or was it a routine backstage aid?
- What did a human specifically edit, verify, or approve?
- What does the publisher’s current policy, contract, or applicable rule require?
A practical approach is to explain routine, low-impact uses in an outlet-wide policy and add a note to an individual work when AI makes a meaningful audience-facing contribution. That is a way to build a consistent policy, not a universal industry requirement.
What major newsroom policies require
These examples show why publishers need their own thresholds. The policies apply to their respective organizations; they do not establish a rule for every journalist, publisher, or country.
#1 Best Overall
| Organization or guidance | What it says |
|---|---|
| Associated Press (AP) | AP’s newsroom standards, updated July 23, 2026, permit listed assistive uses such as early research, document summarization, transcription and translation, headline and summary suggestions, grammar and spelling help, and search optimization. Journalists must review and edit AI output; AI does not replace reporting, sourcing, editorial judgment, or verification. AP requires disclosure when generative AI plays a material role in published content and prohibits generative AI from creating, altering, or enhancing news photography. AP newsroom standards |
| Reuters | Reuters says it retains editorial accountability for all output and will be transparent about how it uses generative AI. Its standards prohibit generative AI from creating or enhancing visual journalism imagery. The policy page does not set one disclosure formula for all text assistance. Reuters AI policy |
| BBC | BBC guidance says generative AI should not directly create news, current affairs, or factual content published or broadcast by BBC News/Nations, except when AI is the subject and its use is illustrative. AI can support production or research under human editorial oversight, and its output must be assessed before use. Certain techniques should be signalled to audiences; synthetic voices not belonging to real people must be clearly disclosed. Cloning a real person’s voice should not normally happen without consent unless there is strong editorial justification and safeguards. BBC guidance on generative AI |
| Impress guidance, as summarized by the UK Parliament | The House of Commons Library’s January 20, 2026 briefing summarizes Impress’s April 2025 best-practice note: publishers should attribute AI use in content generation unless there is a good reason not to. It recommends explaining the tool’s purpose and scope, limitations and risks, and effect on newsgathering, editorial decisions, and published content. It does not prescribe exact wording. The briefing also describes BBC options for direct disclosure at the point of consumption or indirect disclosure outside it. House of Commons Library briefing |
Describe the contribution, not just the tool
A useful disclosure lets readers distinguish limited assistance from generated or altered material. Include only details that accurately describe what happened:
- Task: Say whether AI helped translate, transcribe, summarize, draft, generate an image or audio, or perform another specific task.
- Scope: Identify which part of the published work was affected.
- Editorial effect: Explain relevant effects on reporting, sourcing, editorial choices, or the final material.
- Human review: State what a journalist, editor, translator, or other person actually checked, changed, or approved.
For example, if it is accurate: “We used an AI tool to translate this article from English into Spanish. A member of our Spanish-language staff edited the translation before publication.” This is adaptable wording, not a prescribed AP statement.
Place a content-specific note close enough to the work for readers to know what it refers to, and use accessible language. An outlet policy can cover recurring background practices; a direct note can explain a particular contribution. The right placement depends on the format and the publisher’s policy.
Account for different kinds of AI use
Routine assistance and generated text
Grammar suggestions or a headline option do not necessarily affect a story in the same way as AI-generated passages published with little editing. AP, for example, permits several assistive uses while requiring disclosure when generative AI has a material role in published content. The newsroom should define how it applies that threshold rather than treating every use as equivalent.
Rank #3
Translation
Translation can change a story’s meaning even when the original reporting was human-produced. AP said in a May 8, 2024 explanation that it would clearly indicate AI use in English-to-Spanish translations because translation can materially change a story; it also said AI-generated summaries and suggested headlines were to be edited by journalists. These are AP examples, not rules for all publishers. AP’s explanation of newsroom AI use
Images and audio
Synthetic or altered media raises distinct risks because audiences may take it as documentary evidence. AP and Reuters prohibit generative AI imagery in their respective visual journalism. BBC guidance sets specific disclosure and consent safeguards for synthetic and cloned voices. Follow the applicable outlet policy and label material clearly where required; do not present synthetic media as ordinary reporting material.
Rank #4
- Used Book in Good Condition
Keep accountability and verification explicit
Disclosure does not transfer responsibility to the tool. AP says journalists review and edit AI output and that AI does not replace reporting, sourcing, editorial judgment, or verification. Reuters says it retains editorial accountability for all output.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Check claims and source material independently: AI systems can produce inaccurate or fabricated information. Also consider privacy, information security, copyright, and confidential material before entering it into a tool. BBC guidance flags possible data sharing, use of material for training, and rights concerns.
Best Value
Describe only work that actually happened. Do not say “fact-checked” or “reviewed by an editor” unless those checks occurred. Avoid implying that AI generated an entire piece when it only assisted with one production task, and do not obscure substantial generated material behind a vague statement such as “AI was used.”
What disclosure can—and cannot—promise
A disclosure makes a process more legible, but it is not proof that the content is accurate or trustworthy. A Reuters Institute author reviewing 13 editorial AI guidelines reported that most invoked transparency but often did not explain what it meant in practice. The author also wrote that “we lack any evidence that transparency alone bolsters trust.” The available evidence does not establish a causal effect of disclosure on trust, and labels can affect how people judge accuracy or whether they share content. Do not promise that a label will improve trust. Reuters Institute analysis of AI disclosure frameworks
Build a newsroom policy staff can apply
For recurring uses, an outlet-level policy makes decisions more consistent. The Thomson Reuters Foundation’s April 2025 guide suggests a practical process:
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errors- Inventory the AI tools in use and the tasks they perform.
- Identify risks to journalistic standards and specify how those risks will be mitigated.
- Set up ongoing monitoring and improvement as tools or practices change.
This is policy-development guidance, not a binding standard. Revisit rules as products and newsroom practices change. Policies also differ from legal obligations: these newsroom standards and the UK parliamentary briefing do not settle the law in every country or determine what a publisher contract, academic rule, or platform term requires. Thomson Reuters Foundation responsible AI toolkit
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
