Telegraph Middle East may use software and artificial-intelligence tools to support research organisation, transcription, translation, data processing, editing, production and technical publishing. Human editors remain responsible for everything that is published.
Human editorial responsibility
AI-generated or AI-assisted material must not be published without human review. Editors remain responsible for accuracy, attribution, copyright, fairness, privacy, headline language and the final presentation of the work.
Verification
An AI output is not a source. Facts, quotations, dates, figures and legal or policy claims generated or summarised by a tool must be checked against an identifiable underlying record.
Prohibited uses
AI must not be used to fabricate quotations, sources, events, credentials, awards, employment histories, addresses, social profiles or individual journalist identities. Generated photographs must not be presented as documentary evidence of a real event.
Translation and transcription
Machine translation and transcription may be used for efficiency, but names, numbers, legal language and consequential quotations require review against the original material.
Generated imagery
Illustrative generated imagery should not be used in a way that misleads readers about whether an event occurred or whether a pictured person, location or document is authentic. Material use of generated imagery should be disclosed when necessary for understanding.
Automation and publishing
Automation may assist with metadata, categorisation, image sizing, link checking and distribution. It should not publish unreviewed factual claims automatically. High-risk subjects require additional human oversight.
Corrections
AI-assisted errors are treated in the same way as other errors and are corrected under the Corrections Policy. Responsibility remains with the publication, not the tool.