AI Policy

Artificial intelligence is changing both the technology industry CTRL+PH covers and the tools available to journalists.

CTRL+PH may use artificial intelligence as part of its editorial workflow where it can improve efficiency, organisation, analysis, and clarity. These tools may assist the journalism, but they do not replace editorial judgment, verification, or accountability.

A human remains responsible for every story published by CTRL+PH.

This policy supplements the CTRL+PH Editorial Standards. All AI-assisted editorial work remains subject to those standards.

How we use AI

CTRL+PH may use AI tools to assist with tasks including:

  • organising research and source material;
  • identifying information within documents;
  • summarising source material for editorial review;
  • comparing information across sources;
  • preparing, restructuring, or refining article drafts;
  • improving grammar, structure, and clarity;
  • developing headline, excerpt, and metadata options;
  • formatting information;
  • assisting with calculations or data transformations;
  • helping with routine editorial, publishing, and administrative tasks.

The extent of AI assistance may vary between articles.

AI may contribute substantially to the preparation of a story, but it does not independently determine whether that story should be published.

Human editorial control

AI does not independently publish stories on CTRL+PH.

A human determines what is published, reviews the final article, and remains responsible for its accuracy, sourcing, framing, and presentation.

AI may assist with research, drafting, editing, or production, but decisions involving newsworthiness, context, verification, corrections, and publication remain human editorial responsibilities.

AI is not a source

AI-generated output is not treated as evidence that a factual claim is true.

When AI helps locate, organise, summarise, compare, translate, or analyse information, CTRL+PH should refer back to the original document, statement, interview, database, publication, filing, or other underlying material.

Material produced with AI assistance remains subject to the sourcing and verification requirements in our Editorial Standards.

If reliable verification is unavailable, information should be appropriately qualified or omitted.

Fabrication and hallucinations

CTRL+PH does not knowingly publish information invented by an AI system.

AI tools must not be used to fabricate:

  • facts;
  • quotations;
  • sources;
  • interviews;
  • statistics;
  • documents;
  • events;
  • product specifications;
  • company statements;
  • eyewitness accounts;
  • people or identities.

Plausible-sounding AI output is not sufficient grounds for publication.

Information introduced by an AI system that cannot be traced to a reliable source should either be independently verified or removed.

Quotes

AI tools must not invent, reconstruct, complete, or materially alter quotations.

Quoted statements should be traceable to the underlying source.

Minor formatting or punctuation changes may be made where editorially appropriate, provided the meaning is not changed.

AI-generated approximations of what a person may have meant must never be presented as direct quotations.

Attribution and original work

AI assistance must not obscure where information originated.

AI-generated paraphrasing must not be used to disguise reporting from another publication, avoid attribution, or present another organisation’s work as original CTRL+PH reporting.

Attribution remains governed by the CTRL+PH Editorial Standards.

Confidential and sensitive information

CTRL+PH takes care when using third-party AI services with non-public information.

Embargoed material, confidential communications, source identities, credentials, personal data, unpublished reporting, or other sensitive information should not be submitted to an external AI service unless doing so is appropriate for the circumstances and does not compromise confidentiality, privacy, security, or an agreed embargo.

Protecting sources and sensitive information takes priority over editorial convenience.

AI tools should not be relied upon to determine whether confidential information is safe to disclose.

Images, audio, video, and synthetic media

CTRL+PH will not present AI-generated or materially AI-altered media as an authentic depiction of a real event when it is not.

Synthetic images, video, audio, or other material that could reasonably be mistaken for genuine documentary content should be clearly identified as such.

AI-generated depictions of real people or events must not be used in a way that could reasonably mislead readers about what actually occurred.

Routine design elements, decorative illustrations, or obviously non-documentary graphics may not require the same level of disclosure when their use does not create a misleading impression.

Automated publishing

CTRL+PH does not permit AI systems to autonomously publish editorial content without human review.

Automation may assist with tasks such as formatting, metadata, tagging, scheduling, distribution, or other publishing operations.

The editorial content itself remains subject to human oversight before publication.

Any future increase in publishing automation should preserve meaningful human editorial control.

Institutional bylines

Articles published under the CTRL+PH News Desk byline may have been produced with AI assistance.

The byline represents institutional editorial responsibility. It does not represent an autonomous AI author and does not indicate that an article was automatically published.

The same editorial, verification, sourcing, and corrections standards apply regardless of whether an article carries an institutional or individual byline.

CTRL+PH does not create fictional journalist identities to represent AI systems.

Disclosure

Because AI forms part of CTRL+PH’s normal editorial workflow, routine assistance does not necessarily require a separate disclosure on every article.

Additional disclosure should be provided when AI played a role that is material to understanding how a piece of journalism was produced or that could reasonably affect how a reader evaluates the work.

This may include:

  • substantial AI-generated analysis central to an article;
  • synthetic or materially altered media;
  • AI-assisted data analysis central to the reporting;
  • unusual automated reporting or publishing methods;
  • experiments in AI-generated journalism.

Disclosure should clearly explain the role AI played without exaggerating or obscuring its involvement.

Errors and corrections

Errors in AI-assisted content are CTRL+PH’s responsibility.

The involvement of an AI system does not reduce our responsibility for what we publish and is not an excuse for inaccurate reporting.

Errors are handled under the CTRL+PH Corrections Policy.

Changes to this policy

Artificial intelligence and newsroom practices continue to evolve.

CTRL+PH may update this policy as its editorial workflow, technology, and industry practices change.

Material changes should be reflected on this page.

Responsibility

Technology can assist with research, writing, editing, analysis, and publishing.

It cannot assume editorial accountability.

Our principle is simple: AI may assist with the journalism. A human remains responsible for the journalism.