Newsrooms run on deadlines, and AI runs on prompts. Used well, a chatbot will not replace reporting — but it will happily do the grinding work around it: background research, interview prep, transcript wrangling, headline brainstorming and the million small drafts that eat a journalist’s day. These 30 AI prompts for journalists are ready to copy, organised by newsroom task, each with a one-line note on how to get the most out of it.
One rule before we start: AI is an assistant, never a source. Every fact it gives you must be verified against primary sources before publication. We cover the ethics properly at the end — read it, it matters.
Research & background
Prompt 1:
“Give me a concise backgrounder on [topic or event]: the key players, the timeline of events so far, and the main points of contention. Flag anything that is disputed or uncertain rather than presenting it as fact.”
How to use it: Paste this at the start of a new story to get oriented in minutes, then verify every claim independently.
Prompt 2:
“List the 10 most important primary sources for reporting on [topic] in [country] — official statistics, court records, parliamentary debates, company filings, academic studies — and tell me where to find each one.”
How to use it: Build your source list before you start calling people, so interviews fill gaps instead of repeating basics.
Prompt 3:
“Explain [complex topic, e.g. how the EU AI Act will be enforced] as if briefing a smart reader with no background. Keep it under 300 words and define every piece of jargon.”
How to use it: If you cannot explain it simply, you do not understand it yet — use this as a comprehension test before writing.
Prompt 4:
“What are the strongest arguments for and against [policy or claim]? Present each side’s best case fairly, with the evidence each side cites, without taking a position.”
How to use it: Steelman both sides to avoid strawmanning anyone in your copy.
Interview prep
Prompt 5:
“I am interviewing [name, role] about [topic] next week. Draft 12 questions: 4 warm-up, 4 substantive, 4 challenging follow-ups. Mark which answers I should verify independently afterwards.”
How to use it: Walk into every interview with a plan instead of a blank page.
Prompt 6:
“Here is what [interviewee] has said publicly about [topic]: [paste quotes]. Suggest 5 follow-up questions that probe inconsistencies or push past the talking points.”
How to use it: Essential prep for politicians and PR-trained spokespeople.
Prompt 7:
“Role-play as a defensive press officer for [company or institution]. I will practise my questions, you respond evasively, then critique my interview technique honestly.”
How to use it: Rehearse difficult interviews safely before the real thing.
Prompt 8:
“Draft a professional pre-interview email to a [type of source, e.g. first-time whistleblower] explaining the story, what I need from them, how quotes will be used, and how I will protect their identity.”
How to use it: Turn nervous first-time sources into willing ones.
Article outlines
Prompt 9:
“Turn these reporting notes into a structured outline for a 1,200-word feature: [paste notes]. Suggest a narrative arc, where the strongest quotes should go, and what reporting is still missing.”
How to use it: Beat the blank page after a long week of reporting.
Prompt 10:
“Outline a breaking news story on [event] in inverted-pyramid style: a 30-word lede, the key facts in order of importance, then the context paragraphs a reader needs.”
How to use it: File clean, structured copy fast when the clock is ticking.
Prompt 11:
“Here are excerpts from 15 pages of interview transcripts: [paste]. Identify the 5 strongest narrative threads and suggest which could carry a standalone piece.”
How to use it: Find the story hiding inside your material.
Headline ideas
Prompt 12:
“Write 10 headline options for this story [paste draft or summary]: 3 straight news, 3 with a human angle, 2 data-led, 2 written for social. Keep each under 70 characters.”
How to use it: Give your editor real choices instead of one take-it-or-leave-it headline.
Prompt 13:
“Critique these 5 headlines for accuracy and clickbait risk: [list them]. Rank them and explain which promises the article actually keeps.”
How to use it: Keep headlines honest — reader trust is a journalist’s currency.
Prompt 14:
“Suggest 5 SEO-friendly headlines for [story] that include [keywords] naturally, without distorting what the story actually says.”
How to use it: Serve search readers without selling out the story.
Fact-checking workflows
Prompt 15:
“Here is a draft paragraph containing three factual claims: [paste]. For each claim, tell me what type of primary source would verify it and what search terms to use. Do not verify anything from memory.”
How to use it: Turn AI into a checking assistant — never the authority itself.
Prompt 16:
“List the common statistical traps when reporting on [topic, e.g. crime figures or migration data]: base-rate errors, cherry-picked time periods, correlation vs causation. Give a one-line test for spotting each.”
How to use it: Bulletproof your data paragraphs before the statisticians email you.
Prompt 17:
“Draft a short newsroom policy paragraph on AI-assisted errors: who checks AI output, who signs off, and how we disclose mistakes to readers.”
How to use it: Set the standards before you need them.
Prompt 18:
“I have been sent this viral claim: [paste]. Walk me through a verification workflow step by step: reverse image search, source tracing, timestamp checks, and expert confirmation.”
How to use it: Debunk misinformation methodically instead of reactively.
Transcription summaries
Prompt 19:
“Summarise this interview transcript into three parts: the 3 strongest verbatim quotes, 5 factual claims I must verify, and a 150-word summary. [paste transcript]”
How to use it: Process hour-long interviews in minutes — but always check quotes against the audio before publishing.
Prompt 20:
“Extract every number, date, name and job title mentioned in this transcript into a fact table I can verify line by line. [paste transcript]”
How to use it: Never let a wrong figure slip into print again.
Prompt 21:
“Compare what the interviewee said about [topic] at the start of this transcript versus the end. Flag contradictions, softened positions or walk-backs. [paste transcript]”
How to use it: Catch the moments sources hope you missed.
Data journalism
Prompt 22:
“I have a spreadsheet of [describe the dataset]. Suggest 5 story angles, the calculations each one needs, and the pitfalls to watch for — missing data, outliers, small samples.”
How to use it: Start data stories with questions, not charts.
Prompt 23:
“Write Python/pandas code to clean this dataset: [describe columns and problems]. Explain each step in plain language so I can check the logic.”
How to use it: Do reproducible data work even if you are not a coder — then have someone technical review the code.
Prompt 24:
“Help me write a methodology note for readers explaining how we analysed [dataset], its limitations, and what the data cannot prove.”
How to use it: Transparency in methods builds trust in data stories.
Newsletter writing
Prompt 25:
“Turn these 5 stories into a 300-word newsletter briefing: a strong opening line plus one ‘why it matters’ sentence per story. [paste summaries]”
How to use it: File your morning newsletter in half the time.
Prompt 26:
“Rewrite this newsletter intro in three tones — witty, sober, urgent — without changing a single fact. [paste intro]”
How to use it: Match your newsletter’s voice to the weight of the news.
Prompt 27:
“Suggest 5 subject lines for a newsletter about [topic] that maximise opens without misleading the reader about the contents.”
How to use it: Beat the average open rate without clickbait.
Social promotion
Prompt 28:
“Turn this article into a 5-post thread: a hook, three key points with evidence, and a closing question. Keep each post under 280 characters. [paste article]”
How to use it: Promote stories natively instead of dropping bare links.
Prompt 29:
“Write 3 LinkedIn versions of this story for a professional audience: one data-led, one behind-the-scenes, one with a clear point of view. [paste summary]”
How to use it: Reach the professional readers who never visit homepages.
Prompt 30:
“Draft descriptive alt text for these story images [describe each image], accessible and under 125 characters per image.”
How to use it: Make your visual journalism accessible to every reader.
An honest note on verification and ethics
These prompts will save you hours — but only if you use them with a journalist’s scepticism. AI chatbots hallucinate: they invent quotes, dates, statistics and even court cases with total confidence. Never publish an AI-generated fact without checking it against a primary source. Treat the chatbot like a keen but unreliable intern: fast, helpful, and wrong often enough that you verify everything.
Three more rules. First, protect your sources: never paste unpublished material, source identities or sensitive documents into a public AI chatbot — use your newsroom’s approved, private tools for sensitive work. Second, disclose AI use according to your newsroom’s policy; readers deserve to know how their news was made. Third, mind copyright: AI-generated text sits in a legal grey zone, and the rules differ across Europe — our guide to AI and copyright in the EU explains what creators must know in 2026.
If you also produce content beyond straight news — explainers, video scripts, social formats — our AI prompts for content creators cover that side of the desk too.
The bottom line
Great journalism was never about typing speed — it is about judgment: which story to chase, which question to ask, which fact to check twice. These 30 prompts hand the mechanical work to AI so you can spend your judgment where it counts. Copy the ones that fit your beat, adapt them to your newsroom’s standards, and keep doing the part no chatbot can do: finding things out.