An ethical framework for AI in the newsroom
Combine responsibility, independent verification and a personal rule for AI into an editorial framework a newsroom can explain and uphold. Four principles organise the decision: human accountability, meaningful transparency, authenticity and the protection of the information entrusted to journalists.
The essentials
The human signs and answers. Publication, headlines and framing remain editorial decisions. A tool does not supply your judgment or take responsibility to the audience. Your byline represents a promise whose fulfilment still depends on decisions you understand and evidence you can explain.
Use proportionate transparency. Tell the public when AI has substantially shaped what they receive, such as generated material, synthetic imagery, a cloned voice or the translation of a quotation. The explanation should identify the actual use. Do not present a generated image as a photograph of a news event; label its nature clearly.
Protect data as part of protecting sources. Unpublished interviews, confidential documents and identifying information should not be entered into a tool whose handling of that material you do not control. Consider what a document reveals before uploading it. Editorial responsibility begins during collection and processing, continues through publication and does not disappear because a service performed one step.
Take it to the field
Before sending a document to a tool, identify what it reveals about your sources. Before publication, identify what the audience should know about the AI use. These checks extend the same editorial responsibility.
Your turn
Write a six-point AI policy for your newsroom, or your own freelance practice. Address responsibility, verification, transparency, authenticity and confidential material. Ask two colleagues to review it before adopting it.