ChatGPT Prompts for Journalists

Journalism is under more time pressure than ever — smaller newsrooms, faster news cycles, and audiences that expect depth and speed simultaneously. AI cannot replace the investigative instinct, source relationships, and ethical judgment that define good journalism, but it can dramatically accelerate the research, preparation, and drafting stages of the reporting process. The most important principle when using AI for journalism is verification. Every fact, claim, and quote the AI generates must be independently confirmed. Your prompts should reflect this by instructing the AI to surface questions and angles rather than generate finished copy, to flag any claims it cannot attribute to a specific source, and to present multiple perspectives on contested topics.

Story research prompts should describe the topic, the angle you are considering, and what you already know. Ask the AI to identify background context, relevant data sources, key stakeholders, historical precedents, and potential counterarguments to your thesis. This is faster than starting from a blank search engine query and often surfaces angles you would not have considered. Interview preparation prompts should describe the subject, their role, the story context, and what you need to learn from the conversation. Ask the AI to generate questions organized by topic, starting with rapport-building openers and progressing toward the harder questions. Include follow-up questions for likely responses. Fact-checking prompts should present a specific claim and ask the AI to identify what evidence would be needed to verify or debunk it, suggest where to find that evidence, and flag any common misconceptions about the topic. Headline generation prompts should specify the story angle, the publication's style (tabloid, broadsheet, digital-native), the character limit, and whether you want straight news, feature, or opinion framing. Ask for ten options so you can select or combine the best elements.

Save your journalism prompt templates in PromptingBox and organize them by beat — politics, business, technology, culture. Version your interview prep templates as you refine your questioning technique, and build a library that makes every reporter on your team more thorough and efficient in their research process.

Journalism Prompts You Can Copy and Use Today

Click any prompt to copy it. Replace the {{variables}} with your specific details.

Interview Preparation Brief

Prepare a comprehensive interview brief for my upcoming conversation with {{interviewee_name}}, who is {{interviewee_role_and_context}}.

Story context: {{story_description}}
Angle I am pursuing: {{story_angle}}
What I already know: {{existing_knowledge}}
Interview format: {{format}} (e.g., in-person 30 min, phone 15 min, email Q&A)

Generate:
1. Background research summary — key facts about this person, their public positions on the topic, any previous statements or controversies relevant to the story
2. 15-20 interview questions organized in this order:
   - 2-3 rapport-building openers (genuine, not generic)
   - 5-6 core questions that advance the story angle
   - 3-4 follow-up questions for likely responses or deflections
   - 2-3 accountability questions (harder questions saved for when rapport is established)
   - 1 closing question that gives them a chance to add anything
3. Potential landmines — topics they are likely to deflect on, and suggested redirect strategies
4. Key documents or data points to reference during the interview
5. Quotes from their previous public statements that I can ask them to expand on or reconcile

Flag any claims in your background research that I should independently verify before the interview. Do NOT fabricate quotes or attribute statements you are not certain about — mark uncertain items clearly.
interviewee_nameinterviewee_role_and_contextstory_descriptionstory_angleexisting_knowledgeformat

Why it works: Organizing questions by conversation arc (rapport to accountability) mirrors proven interviewing technique. Flagging unverified claims prevents the journalist from citing AI-generated misinformation.

Article Outline and Structure

Create a detailed article outline for the following story:

Topic: {{topic}}
Angle: {{angle}}
Publication: {{publication}} (style: {{publication_style}})
Target length: {{word_count}} words
Deadline: {{deadline}}
Audience: {{audience}}

Sources I have:
{{available_sources}}

Key facts established so far:
{{key_facts}}

Generate:
1. A working headline and 2 alternative options
2. A lede (opening paragraph) — write 3 options:
   - Anecdotal lede (scene-setting)
   - Summary lede (hard news style)
   - Question lede (provocative opening)
3. Section-by-section outline with:
   - Section heading / purpose
   - Key points to cover
   - Which source supports each point
   - Suggested transitions between sections
4. A "nut graf" — the paragraph that tells the reader why this story matters now
5. Suggested kicker (closing) — 2 options (circular return to lede, or forward-looking)
6. Gaps analysis — what information is still missing and what sources could fill those gaps
7. Potential counterarguments or opposing viewpoints to address for balance

Structure this as a {{article_type}} (e.g., news analysis, feature, investigative, profile, explainer).
topicanglepublicationpublication_styleword_countdeadlineaudienceavailable_sourceskey_factsarticle_type

Why it works: Offering three lede options gives the writer creative choices rather than locking into one approach. The gaps analysis turns the outline into a reporting roadmap, not just a writing plan.

Fact-Check Verification Plan

Create a fact-check verification plan for the following claims in my article:

Article topic: {{article_topic}}
Publication date: {{publication_date}}

Claims to verify:
{{claims_list}}

For each claim, provide:
1. The specific assertion being made
2. What type of evidence would confirm or debunk it (document, data, expert testimony, public record)
3. Three specific sources to check (databases, government records, academic papers, organizations) — include URLs or access instructions where possible
4. Common misconceptions or misinterpretations related to this claim
5. How to determine if the claim is: fully verified, partially true, misleading, or unverifiable
6. The confidence level of your own assessment (high / medium / low) and why

Also generate:
- A master checklist I can print and work through
- Red flags to watch for (cherry-picked statistics, correlation-causation errors, outdated data presented as current)
- Suggested language for the article if a claim cannot be fully verified (attribution, hedging, transparency)

IMPORTANT: If you are unsure about any factual claim, say so explicitly. Do not present uncertain information as verified. Mark any items where your training data may be outdated.
article_topicpublication_dateclaims_list

Why it works: Requiring specific evidence types and multiple source suggestions creates a systematic verification process. The explicit instruction to flag uncertainty prevents the AI from confidently presenting unverified information.

Source Evaluation Matrix

Evaluate the following sources for my story on {{story_topic}}. For each source, assess their credibility and usefulness for the article.

Sources to evaluate:
{{sources_list}}

For each source, analyze:
1. Credibility assessment:
   - Expertise: What qualifies them to speak on this topic?
   - Potential bias: What interests, affiliations, or funding could influence their perspective?
   - Track record: Have they been reliable in previous reporting? Any retractions or controversies?
   - Access: Do they have direct knowledge or are they interpreting secondhand information?
2. Usefulness rating (1-5): How central is this source to the story?
3. What this source can provide that no other source can
4. What this source is likely to say vs. what I need them to confirm
5. Risks of relying on this source (legal, reputational, accuracy)
6. Corroboration needed: What other sources should confirm key claims from this source?

Then provide:
- A source map showing which sources corroborate each other and where there are conflicts
- Gaps: Perspectives or expertise missing from my source list
- Suggested additional sources to seek (by role/expertise, not specific names if you cannot verify they exist)
- An on-the-record vs. background recommendation for each source

Story context: {{story_context}}
Story angle: {{story_angle}}
Publication type: {{publication_type}}
story_topicsources_liststory_contextstory_anglepublication_type

Why it works: Analyzing bias and corroboration needs for each source prevents over-reliance on a single perspective. The source map reveals where the reporting is thin before the story publishes.

Headline Generator and A/B Options

Generate headline options for the following article:

Story summary: {{story_summary}}
Key finding or news: {{key_finding}}
Publication: {{publication_name}}
Publication style: {{style}} (e.g., tabloid, broadsheet, digital-native, magazine)
Section: {{section}} (e.g., front page, business, opinion, features)
Character limit: {{character_limit}}
SEO keyword (if digital): {{seo_keyword}}

Generate 10 headlines in these categories:
1. Straight news (2 options) — who did what, clear and direct
2. Feature/narrative (2 options) — evocative, draws the reader into a story
3. Question format (2 options) — poses the central question the story answers
4. Data-driven (2 options) — leads with the most striking number or statistic
5. Provocative/opinion (2 options) — takes a stance or challenges assumptions

For each headline:
- Note the word count and character count
- Rate its clickability (1-5) and clarity (1-5) — a headline can be clickable but unclear, which is a problem
- Suggest a matching subheadline / deck (1-2 sentences) that complements it
- Flag if it could be perceived as misleading or sensationalized

Then recommend your top 3 picks and explain why they would perform best for this publication's audience.
story_summarykey_findingpublication_namestylesectioncharacter_limitseo_keyword

Why it works: Categorizing by headline type ensures variety rather than 10 variations of the same framing. Rating clickability AND clarity separately prevents the common trap of optimizing for clicks at the expense of accuracy.

Investigative Lead Tracker

Help me organize and prioritize investigative leads for my ongoing story:

Investigation topic: {{investigation_topic}}
Working thesis: {{working_thesis}}
Time invested so far: {{time_invested}}
Resources available: {{resources}} (e.g., FOIA budget, travel budget, team size)

Current leads and tips:
{{leads_list}}

For each lead, analyze and categorize:
1. Priority level (high / medium / low) based on:
   - Potential impact if the lead pans out
   - Likelihood of yielding usable information
   - Resource investment required (time, money, legal risk)
2. Next action — the single most important step to advance this lead
3. Required resources (FOIA request, travel, database access, legal review)
4. Timeline estimate to develop this lead
5. What this lead could prove or disprove in the larger story
6. Dead-end indicators — signs that it is time to stop pursuing this thread

Also provide:
- A recommended investigation sequence (which leads to pursue first and why)
- Connections between leads that might indicate a larger pattern
- Public records and databases to search for corroborating evidence
- Legal considerations (shield law applicability, defamation risk, source protection)
- A decision framework for when to publish what I have vs. continue investigating

Assume I am an experienced reporter — focus on strategic advice, not basic journalism instructions.
investigation_topicworking_thesistime_investedresourcesleads_list

Why it works: Prioritizing leads by impact AND resource cost prevents the common investigative trap of spending weeks on a lead that yields a minor detail. The dead-end indicators save time by defining when to move on.