Critical Registrar may use artificial intelligence tools in parts of our workflow. This page explains our principles for that use and where the line sits between assistance and authorship.

Our Principles

Accuracy comes first. Nothing about using AI changes our standard for what we publish. Any output produced with AI assistance is held to the same verification requirements set out in our Fact-Checking policy before it runs.

A person is always in charge. AI tools support our journalists; they do not replace editorial judgment. A reporter or editor reviews and takes responsibility for anything AI has touched before it is published, the same way they would for a colleague's draft.

We say when AI was involved. Where AI has played a substantial role in producing something readers see, whether that's an illustration, a summary, or a translated version of a story, we disclose it. Readers should never have to guess whether they're looking at AI-assisted content.

Independence over convenience. We evaluate any AI tool or vendor we use, whether built in-house or licensed, against the same standards of accuracy and fairness that govern the rest of our reporting. Speed or cost never override that evaluation.

Fair representation matters. We're attentive to bias in AI-generated text and images, including how AI tools depict people, cultures, and perspectives, particularly in coverage of Africa, where AI models are more prone to error and stereotype.

Our journalism is our own. The reporting, analysis, and writing that define Critical Registrar are the product of our journalists' work. We use AI to support that work, not to generate it wholesale, and we take steps to protect our published journalism from being used to train AI systems without our consent.

What We Use AI For

Our journalists may use AI tools to assist with research, translation, transcription, data analysis, and early drafts of routine material such as headlines or story summaries. In every case, a journalist reviews the result for accuracy, fairness, and tone before it's published, and edits or discards anything that doesn't meet our standards.

Any use of AI beyond these routine tasks, such as AI-generated images used in place of photography, or AI-written passages appearing substantially as written, requires editor approval and is clearly labeled for readers.

Questions or Concerns

If you have a question about how AI was used in a specific piece, or believe a disclosure is missing, contact us.