AI Policy

Last updated

FinanceInTime uses artificial intelligence (AI) tools in its newsroom. This page explains what those tools do, what they do not do, and how our editors remain responsible for everything we publish.

In short

  • AI tools help us monitor sources, gather research and prepare first drafts.
  • No article is published by an AI tool. Every article is reviewed, edited, fact-checked against its sources and approved by a named human editor.
  • That editor is shown as the byline and holds editorial responsibility for the article.
  • Opinion articles are written by people, not by AI.
  • AI-generated images are used only as generic illustrations, and they are labelled.
  • We do not use our readers' personal data to train AI.

What AI tools do in our newsroom

We currently use large language model tools (at present, Claude, made by Anthropic) for these tasks:

  • Monitoring sources. Following official sources such as central banks, statistics agencies, regulators, exchanges and company announcements, and spotting new releases and developments.
  • Research. Collecting the relevant documents and data, earlier readings and historical comparisons, and the official calendar of upcoming events.
  • First drafts. Preparing a first draft of a news article, brief or analysis in our house style. Each draft comes with the list of sources it is based on and a list of its factual claims, each linked to the passage of the source that supports it.
  • Verification support. A separate check compares the claims in a draft with their sources and flags anything that does not match, for the editor to resolve.
  • Avoiding repetition. Checking whether a development has already been covered, so that new information updates an existing article instead of producing a duplicate.

What our editors do

Every category on FinanceInTime has a named editor, listed on our masthead. Before anything is published, the editor responsible for the category:

  1. reads the whole draft from start to finish and edits it: they rewrite, add, cut or reject material as needed;
  2. checks every figure, name and date against the linked sources, and resolves every claim the verification check has flagged;
  3. confirms that the sources are listed, with the primary source cited where one exists;
  4. confirms that the headline and summary are accurate and not sensational, and that the article contains no investment advice;
  5. confirms that any image is licensed for the use and credited;
  6. approves the article for publication.

The editor confirms each of these steps on a checklist that is recorded with the article. Our publishing system keeps every version of the text and a log of who changed and approved it. AI tools in our newsroom can prepare drafts only: they cannot publish an article or change a published one.

Editors also decide what not to publish. Many drafts are rejected because they are not newsworthy enough, duplicate earlier coverage or cannot be verified.

Bylines and responsibility

The byline on an article names the editor who reviewed and approved it for publication (or, in some cases, another member of our staff who wrote or substantially reworked it). That person is responsible for the accuracy and fairness of the article, and our editorial standards apply to it in full, whatever tools were used to prepare it. Bylines link to the author's profile.

What we do not use AI for

  • Opinion. Opinion articles express the views of their named author and are written by that person.
  • Quotes. AI tools are never used to create or alter quotes. Every quote comes from a source and is checked against it.
  • Images of real people or events. We do not publish AI-generated images that depict real people, real events or real company logos, or that could be mistaken for photographs of news events. When we use an AI-generated illustration of a general theme (for example an abstract image representing markets), its credit says that it is an AI-generated illustration.
  • Investment advice. Our articles, however they are prepared, do not give personal investment advice.
  • Publishing decisions. What we publish, and when, is decided by our editors.

Your data

We do not use readers' personal data, such as newsletter email addresses, to train AI models, and we do not enter it into AI tools. The AI tools we use work with public sources and with our own drafts. See our privacy policy.

Mistakes

AI tools can make mistakes, which is one reason why every article is checked by an editor. Editors can make mistakes too. If you find an error in any of our articles, please tell us at corrections@financeintime.com. Our corrections policy explains how we fix and label errors.

Changes to this policy

We review this policy as our tools and practices change, and we will update this page when they do. The date at the top shows when it was last changed.

Contact

Questions about how we use AI: editorial@financeintime.com.