Writing an AI policy for your newsroom: what to include
Using AI without a policy invites mistakes. These are the points most newsroom AI policies cover, in plain language.
More newsrooms now publish a short policy explaining how they use AI. It protects your audience’s trust and gives your team clear rules.
What AI may be used for
For example: first drafts from the reporter’s notes, headlines, translation, transcription, summaries of documents.
What it must not do
For example: invent quotes, create images of real events, publish without human review, or be the only source for a fact.
Human review
Say who checks AI-assisted work before publication, and what they check: facts, names, numbers, quotes and fairness.
Disclosure
Decide when you tell readers that AI was used, and how.
Data and privacy
Say what material may be put into AI tools. Confidential sources and unpublished investigations may need extra rules.
Keep it short and review it
A one-page policy that people read beats a long one nobody does. Review it every few months as tools change.
See how Byline approaches this on our Responsible AI page.