Skip to main content
Back to Blog
journalism promptsAI for journalistsfact-checkinginvestigative journalismheadline writinginterview prep

50 AI Prompts for Journalists: Research, Write, and Verify Faster

AI prompts for journalists covering interview prep, fact-checking, story angles, headline writing, and investigative research. Copy-paste templates.

SurePrompts Team
April 1, 2026
22 min read

AI prompts that help journalists research deeper, write tighter, and verify faster. Every template is ready to copy and use.

82% of Journalists Now Use AI — Most Use It Badly

Muck Rack's 2026 State of Journalism report is clear. 82% of journalists now use at least one AI tool, up from 66% last year. ChatGPT leads at 47%.

But using AI and using it well are different things.

82%
Of journalists now use at least one AI tool, per Muck Rack's 2026 State of Journalism report

The Reuters Institute's 2026 Trends report found that 64% of news executives consider back-end automation their most important AI use case. Yet most journalists still type half-formed questions and get mediocre results.

These 50 prompts change that. They cover interview prep, fact-checking, story angles, headlines, investigative research, and source analysis. Each prompt is built for journalism-specific workflows.

Build custom prompts for your beat with the AI prompt generator.

The Journalism Prompt Framework

Journalists need different things from AI than other professionals. Accuracy matters more than speed. Verification matters more than volume.

Info

The Journalism Prompt Framework: Beat Context + Specific Task + Source Requirements + Ethical Constraints + Output Format. This structure keeps AI aligned with journalistic standards.

Here is a bad prompt versus a good one.

Before

Write me a story about housing prices.

After

Act as a research assistant for a business reporter covering U.S. housing. Analyze the current housing affordability crisis in [city]. Identify the three most significant data points from government sources (Census Bureau, HUD, local housing authority). Provide specific numbers with source citations. Do not generate any statistics — only use verifiable data I can confirm.

The good prompt sets boundaries. It demands sources. It prohibits fabrication. That is what journalism requires.

Interview Preparation Prompts (1–10)

Great interviews start with great preparation. AI cuts research time without cutting depth.

According to the Reuters Institute, 22% of UK journalists use AI monthly for story research. These prompts make that research more rigorous.

1. Source Background Briefing

code
Act as a research assistant. I'm interviewing [name, title, organization] about [topic]. Compile a background briefing: their professional history, published positions on [topic], any controversies or notable public statements, recent organizational developments, and connections to other key figures in this space. Cite all sources by name. Flag any claims you cannot verify.

2. Interview Question Generator

code
Act as a senior editor. I'm interviewing [person/role] about [topic] for a [publication type]. Generate 15 interview questions organized by: opening questions (rapport building), core topic questions (news value), follow-up probes (for likely responses), accountability questions (challenging claims), and closing questions (forward-looking). Avoid yes/no questions. Each question should seek specific facts or commitments.

3. Pre-Interview Fact Sheet

code
Act as a fact-checker. Create a fact sheet for my interview with [subject] about [topic]. Include: key statistics with named sources, timeline of relevant events, list of stakeholders and their positions, common claims about this topic (with verification status), and specific data points I can use to challenge vague assertions. Mark any unverified claims clearly.

4. Expert Source Identification

code
Act as a research assistant. I'm reporting on [topic/issue]. Identify 10 potential expert sources across these categories: academic researchers (with relevant published work), government officials (with jurisdiction), industry practitioners, advocacy groups (both sides), and affected community members. For each, provide their expertise area and why they matter to this story. Focus on sources who have public contact information.

5. Hostile Interview Preparation

code
Act as a media trainer (from the journalist's perspective). I'm interviewing [public figure/official] about [controversial topic]. They are likely to: deflect, use talking points, or give non-answers. For each of my 5 key questions below, generate 3 anticipated deflections and a follow-up probe that redirects to the original question. [List your 5 questions.]

Tip

Always verify AI-generated background information before using it in an interview. AI models can conflate people with similar names or generate plausible but inaccurate biographical details.

6. Data Interview Preparation

code
Act as a data journalist. I'm interviewing [official/expert] about [dataset or report]. Help me prepare by: identifying the 5 most newsworthy findings in this data, calculating key comparisons (year-over-year, peer comparisons), flagging methodological limitations I should ask about, and drafting 5 questions that use specific numbers to hold the source accountable.

7. Cross-Reference Source Claims

code
Act as a research assistant. My source claims: [paste specific claims]. For each claim, find: publicly available evidence supporting it, publicly available evidence contradicting it, the original source of the claim if it's been repeated elsewhere, and key context that changes the interpretation. Be explicit about what you cannot verify. Do not generate fake sources.

8. Cultural Context Briefing

code
Act as an area specialist. I'm reporting on [topic] in [country/region/community]. Brief me on: relevant cultural context that affects interpretation, historical background essential for accurate reporting, key local figures and power structures, common Western misconceptions about this topic, and terminology I should use or avoid. Cite named sources for factual claims.

9. Follow-Up Interview Strategy

code
Act as a senior reporter. My first interview with [source] about [topic] produced these key quotes and claims: [paste notes]. Identify: claims that need verification, contradictions with public records, areas they avoided or deflected, new questions raised by their answers, and other sources I should contact to corroborate or challenge their account.

10. Panel Discussion Preparation

code
Act as a producer. I'm moderating a panel on [topic] with [X panelists from Y backgrounds]. Create a discussion framework: opening question that engages all panelists, topic blocks with transition questions, provocative questions that create productive disagreement, audience question themes to anticipate, and time management cues for a [X-minute] discussion.

Fact-Checking and Verification Prompts (11–20)

Verification separates journalism from content. AI accelerates the checking process.

According to the Reuters Institute's 2026 conference, fact-checked claims involving AI-generated content doubled year-over-year — from 7% to 16% in Brazil alone. Verification has never been more important.

11. Claim Verification Checklist

code
Act as a fact-checker at a major news organization. Evaluate this claim: "[paste claim]." For each element of the claim: identify what can be verified against public records, list specific databases or sources to check, note what evidence would confirm or refute it, flag the difference between verified facts and reasonable inferences, and assess the overall reliability. Rate confidence: confirmed, likely true, unverified, likely false, or false.

12. Statistical Claim Audit

code
Act as a data journalist. Audit this statistical claim: "[claim with numbers]." Check: Is the original study or dataset identifiable? What was the sample size and methodology? Is the statistic being used in its original context? Are there more recent data points? What are common misinterpretations of this type of statistic? Provide the correct interpretation with caveats.

13. Image and Video Verification Guide

code
Act as a digital verification specialist. I need to verify this [image/video] that claims to show [description]. Walk me through a verification protocol: reverse image search strategy, metadata analysis points to check, geolocation verification steps using visual landmarks, weather/lighting consistency checks, and social media provenance tracing. List specific free tools for each step.

16%
Of fact-checked claims in Brazil involved AI-generated content in 2025, double the prior year, per fact-checker Aos Fatos

14. Document Authenticity Check

code
Act as an investigative researcher. I've received a [document type: memo/email/report] allegedly from [source/organization] dated [date]. Create a verification checklist: formatting and style consistency with known documents from this source, terminology and jargon accuracy, timeline consistency with known events, metadata points to examine, and independent sources who could confirm authenticity.

15. Source Credibility Assessment

code
Act as a media literacy expert. Assess the credibility of this source: [organization/website/individual] who claims [statement]. Evaluate: track record of accuracy, funding sources and potential conflicts of interest, editorial standards or correction policies, expertise relevant to the claim, and how other credible sources reference them. Distinguish between the source's overall credibility and this specific claim's reliability.

16. Viral Content Debunking

code
Act as a fact-checker. This claim is going viral: "[paste viral claim]." Create a debunking structure: what exactly is being claimed (steel-man the strongest version), what verifiable facts exist, what context is missing from the viral version, what the original source actually said versus how it was characterized, and a clear plain-language correction. Keep the correction under 150 words.

17. Timeline Verification

code
Act as a chronology specialist. Verify this timeline of events: [paste claimed sequence of events]. For each event: confirm the date against public records, identify gaps or inconsistencies, check whether the causal chain is supported or assumed, flag events that cannot be independently verified, and note any events that are missing but relevant.

18. Financial Disclosure Analysis

code
Act as a financial journalist. Analyze this financial disclosure: [describe document — campaign filing, corporate report, nonprofit 990]. Identify: unusual patterns in revenue or spending, related-party transactions, discrepancies with public statements, comparison with peer organizations, and specific line items that warrant further reporting. Flag items that may require FOIA requests for clarification.

19. Scientific Study Evaluation

code
Act as a science journalist. Evaluate this study: [paste title, journal, key findings]. Assess: study design (sample size, methodology, control groups), statistical significance and effect size, funding source and potential conflicts, whether conclusions match the data, how it fits within existing literature (does it confirm or contradict consensus?), and limitations the authors acknowledge versus limitations they don't.

20. Public Records Request Draft

code
Act as an investigative reporter. Draft a public records request (FOIA/state equivalent) for [specific records] from [government agency]. Include: precise description of records sought, relevant date range, legal basis for the request citing specific statutes, format preference (electronic), fee waiver justification (public interest), and expedited processing argument if applicable. Make it specific enough to avoid overbreadth objections.

Story Development Prompts (21–32)

Finding the angle is half the battle. AI helps identify gaps competitors miss.

Medianet's 2026 Australian Media Landscape report found that 54% of journalists now use AI in their work. Those who use it for story development — not drafting — gain the most competitive advantage.

21. Story Angle Generator

code
Act as an enterprise editor. I have this news event: [describe event/development]. Generate 10 story angles organized by: the immediate news angle, the human impact angle, the systemic/policy angle, the historical context angle, the data/investigation angle, the follow-up angle (what happens next), and the contrarian angle (what everyone is getting wrong). For each, suggest a specific headline and primary source type.

22. Beat Trend Analysis

code
Act as a research analyst for a [beat] reporter. Analyze emerging trends in [topic area] over the past 6 months. Identify: patterns in recent developments that suggest a larger story, data sources that could quantify the trend, affected populations or stakeholders, policy responses (or lack thereof), and potential future developments. Cite specific recent events that illustrate each trend.

23. Investigative Story Outline

code
Act as an investigative editor. I'm pursuing a story about [topic/allegation]. Create an investigation plan: hypothesis to test, key documents to obtain (with specific FOIA targets), source categories to develop (insiders, victims, experts, officials), data analysis opportunities, legal considerations, and a publication timeline with milestones. What would make this story bulletproof?

24. Enterprise Story Pitch

code
Act as a features editor evaluating pitches. Help me develop this story idea into a pitch: [describe idea]. Structure the pitch as: the news peg (why now), the central question or tension, what reporting will reveal that isn't already known, proposed sources (3 minimum), estimated scope and timeline, and the so-what for readers. Keep the pitch under 300 words.

Warning

Never use AI to generate quotes, fabricate sources, or create fictional scenarios presented as fact. The Reuters Institute documented cases in 2025 where AI-generated journalism slipped past editors undetected. Always verify and attribute.

25. Data Story Identification

code
Act as a data editor. Here is a dataset about [topic]: [describe dataset, variables, size]. Identify the 5 most newsworthy findings by looking for: outliers and anomalies, trends over time, geographic patterns, demographic disparities, comparisons with benchmarks or standards, and correlations worth investigating. For each finding, suggest a headline and the verification steps needed.

26. Localization Template

code
Act as a local reporter. A national story about [topic] is breaking. Help me localize it for [city/region]. Identify: local data equivalents for national statistics, local officials with jurisdiction over this issue, affected communities or institutions in our coverage area, local policy context that differs from the national picture, and local sources who can provide on-the-ground perspective.

27. Explanatory Journalism Framework

code
Act as an explainer editor. I need to help readers understand [complex topic]. Create an explanatory framework: the core concept in one sentence (no jargon), the 3 most common misconceptions, an analogy that makes it intuitive, the key stakeholders and their interests, what changed recently to make this newsworthy, and specific reader actions or implications. Target a general audience reading level.

28. Solutions Journalism Approach

code
Act as a solutions journalism specialist. I'm reporting on [problem]. Help me identify potential solutions angles: organizations or communities that have measurably reduced this problem, specific interventions with documented outcomes, transferability analysis (would it work elsewhere?), limitations and critiques of the solutions, and evidence-based metrics for evaluating success. Cite named examples.

29. Accountability Story Framework

code
Act as an accountability reporter. [Entity] promised [specific commitment] on [date]. Create a framework for measuring whether they delivered: specific benchmarks from the original commitment, data sources to measure progress, stakeholder perspectives on results, comparison with peer organizations' performance, and a scorecard format for presenting findings.

30. Seasonal Story Calendar

code
Act as a planning editor. Create a 12-month story calendar for my [beat]. For each month, suggest: a data-driven story tied to regular releases (government reports, quarterly earnings, annual statistics), an enterprise story opportunity tied to seasonal patterns, a profile or feature opportunity, and relevant anniversary hooks. Focus on stories that require advance planning and records requests.

31. Series Architecture

code
Act as a projects editor. I'm planning a multi-part series on [topic]. Design the architecture: how many parts (recommend optimal number), what each installment covers, the narrative arc across the series, data visualization opportunities per installment, multimedia elements to consider, and publication cadence recommendation. Each part must stand alone while building toward the series thesis.

32. Reader Engagement Story Design

code
Act as an audience engagement editor. I'm publishing a story about [topic]. Design an engagement strategy: questions to pose to readers before publication (callout), interactive elements within the story, social media discussion starters, follow-up story opportunities based on reader response, and specific communities or groups to share it with. Focus on generating meaningful reader contributions for future reporting.

Headline and Writing Prompts (33–42)

Headlines drive traffic. Clear writing builds trust. AI helps with both.

The Reuters Institute found that 16% of UK journalists use AI at least monthly for generating parts of text articles like headlines. These prompts focus on the writing tasks where AI adds the most value.

33. Headline A/B Testing

code
Act as a headline writer for a [publication type]. Write 10 headline options for this story: [one-paragraph summary]. Create variations across these styles: straightforward news headline, number/data-driven headline, question headline, how/why headline, and contrast/tension headline. Each must be under 70 characters. Each must be factually accurate. Rank them by likely click-through rate.

34. Lead Paragraph Options

code
Act as a writing coach. I'm writing a [story type] about [topic]. Write 5 different lead paragraph options: a hard news lead (who/what/when/where), an anecdotal lead (scene-setting), a data lead (striking statistic), a quote lead (compelling source quote I'll supply), and a contrast lead (then vs. now). Each lead should be under 40 words. Each must hook the reader immediately.

35. Story Structure Template

code
Act as a narrative editor. Help me structure this story: [summarize key facts, sources, data]. Recommend the best structure from: inverted pyramid, narrative chronology, nut graf, hourglass, or diamond. Explain why that structure fits this material. Then provide a detailed outline with section transitions. Indicate where to place quotes, data, and context.

36. Complex Topic Simplification

code
Act as an explainer writer. Simplify this technical concept for a general audience: [paste technical content or jargon-heavy text]. Rewrite it using: no jargon without immediate definition, sentences under 20 words, concrete examples instead of abstractions, and analogies where helpful. Maintain factual accuracy. Flag any simplification that sacrifices important nuance.

Tip

Use AI for headline variations and structural suggestions, not for writing your story. The journalist's voice, judgment, and sourcing are what readers trust. AI is your research assistant, not your ghostwriter.

37. Nut Graf Workshop

code
Act as a writing editor. Here is my story so far: [paste first few paragraphs and key facts]. Write 5 options for the nut graf — the paragraph that tells readers why this story matters now and what it's about. Each should: connect the specific to the universal, establish stakes, and signal the story's direction. Keep each under 50 words.

38. Transition Writing

code
Act as a copy editor. I need transitions between these story sections: [describe or paste section summaries]. Write 3 options for each transition that: move the reader smoothly between ideas, avoid formulaic phrases ("meanwhile," "on the other hand"), maintain narrative momentum, and don't introduce new information. Each transition should be one sentence.

39. Data Visualization Narrative

code
Act as a data visualization editor. I have this data: [describe dataset and key findings]. Write the narrative text that should accompany a chart/graph. Include: a headline for the visualization, the key takeaway in one sentence, context that the visual alone doesn't convey, methodology note, and source attribution. Keep total text under 100 words.

40. Social Media Summary

code
Act as a social media editor. Write social media posts promoting this story: [one-paragraph summary]. Create: a Twitter/X post (under 280 characters with a hook), a LinkedIn post (3-4 sentences, professional tone), an Instagram caption (conversational, with hashtag suggestions), and a newsletter blurb (2 sentences highlighting the key finding). Each must accurately represent the story without sensationalizing.

41. Editor Memo

code
Act as a reporter writing to your editor. Draft a memo explaining why we should invest resources in this story: [describe story idea]. Include: news value assessment, competitive landscape (has anyone covered this?), source access and feasibility, required resources (time, travel, FOIA costs), potential legal risks, and expected impact. Be honest about challenges. Keep it under 500 words.

42. Correction and Clarification Draft

code
Act as a standards editor. Draft a correction for this error: [describe what was published and what was wrong]. Follow journalism correction standards: state what was published, state what is correct, explain the nature of the error (factual, contextual, attribution), and note when the article was updated. Be direct and transparent. Do not minimize or explain away the error.

Source Analysis and Research Prompts (43–50)

Deep research separates good journalism from great journalism. AI extends your reach.

43. Corporate Structure Analysis

code
Act as a business researcher. Map the corporate structure of [company/organization]. Identify: parent company and subsidiaries, key executives and board members, major shareholders or funders, recent acquisitions or divestitures, regulatory filings (SEC, state registrations), and connections to other entities relevant to my story about [topic]. Use only publicly available information. Cite each source.

44. Lobbying and Influence Mapping

code
Act as a political researcher. Map the lobbying and influence network around [policy issue/legislation]. Identify: registered lobbyists and their clients (from public disclosures), campaign contributions from affected industries, revolving door connections (officials who joined industry or vice versa), industry trade groups and their public positions, and think tanks producing relevant research (with their funders). Cite disclosure databases.

45. Court Records Research Guide

code
Act as a legal researcher. I'm investigating [person/organization] for a story about [topic]. Guide me through finding relevant court records: federal (PACER/RECAP) and state court databases to search, case types most relevant to this investigation, key search terms and name variations to try, how to read docket entries efficiently, and what filings to prioritize (complaints, depositions, settlements). Note which databases are free vs. paid.

46. Government Budget Analysis

code
Act as a fiscal analyst. Analyze this government budget document for [agency/jurisdiction]: [describe budget or paste key figures]. Identify: largest line-item changes year-over-year (increases and decreases above 10%), items that don't match public statements or priorities, hidden or obscured spending, comparison with peer jurisdictions, and unfunded mandates or deferred costs. Present findings as a summary memo with specific dollar amounts.

47. Regulatory Filing Review

code
Act as a regulatory analyst. Review this filing: [describe document — SEC 10-K, FDA submission, EPA permit, FCC application]. Identify: key disclosures that could be newsworthy, risk factors that weren't previously public, changes from prior filings, technical language that obscures important information, and specific items that warrant further reporting or expert comment.

TaskAI StrengthHuman Required
Background researchStrong pattern matchingSource verification
Data analysisFast calculationInterpretation and context
Headline writingVolume and variationEditorial judgment
Fact-checkingCross-reference speedFinal verification
Source identificationBroad scanningRelationship building
Interview prepComprehensive briefingsReading the room

48. Nonprofit Investigation Starter

code
Act as an investigative researcher. I'm looking into [nonprofit organization]. Guide me through publicly available data: IRS Form 990 analysis (key ratios to examine, red flags), state charity registration filings, board member connections and potential conflicts, executive compensation benchmarks, program spending vs. fundraising ratios, and any enforcement actions. Recommend specific databases and tools.

49. International Source Verification

code
Act as a foreign affairs researcher. I'm verifying claims from a source about [topic] in [country]. Help me navigate: relevant government databases and registries in that country, local media outlets to cross-reference, international organizations with relevant data (UN, World Bank, regional bodies), academic researchers specializing in this area, and cultural or linguistic considerations that affect interpretation. Note translation tools and their limitations.

50. Story Archive and Follow-Up Research

code
Act as a research librarian. My publication covered [topic] extensively [timeframe]. Help me identify follow-up opportunities: compile a checklist of promises, deadlines, and commitments from our past coverage that are now due, identify stakeholders who should be re-contacted, find new data releases that update our previous reporting, and suggest anniversary hooks or milestone moments for revisiting the story.

Ethical Guidelines for AI in Journalism

The Medianet 2026 Australian Media report found that 93% of journalists express concern about AI's impact on journalistic integrity. These concerns are valid.

1

Never present AI-generated text as original reporting. AI assists your work — it does not do your work.

2

Verify every fact, quote, and statistic AI provides. Models hallucinate plausible-sounding details.

3

Disclose AI use per your organization's policy. Transparency builds reader trust.

4

Never use AI to generate fake quotes, fabricate sources, or create misleading content.

5

Keep AI out of final editorial decisions. News judgment requires human accountability.

The Reuters Institute found only 12% of the public is comfortable with news made entirely by AI. Audiences trust human journalists. AI should strengthen that trust, not undermine it.

Learn more about structured prompting in our prompt engineering basics guide. For terminology, see our prompt engineering glossary. Writers and content creators can also explore AI prompts for writers.

Frequently Asked Questions

Is it ethical for journalists to use AI?

Yes, when used as a research and productivity tool with full transparency. The key is human oversight. AI assists with research, structure, and verification — not reporting or writing the final story. Check your newsroom's AI policy.

Which AI tool is best for journalists?

ChatGPT and Claude both work well for research tasks. Claude handles longer documents and nuanced analysis effectively. Use the ChatGPT prompt generator to build prompts optimized for journalism workflows.

Can AI fact-check articles?

AI can cross-reference claims against known information and flag inconsistencies. But it cannot access proprietary databases, make phone calls, or verify claims through human sources. AI is a first pass, not a final check.

Will AI replace journalists?

No. The Reuters Institute's 2025 report found that only 12% of the public is comfortable with AI-only news. Journalism requires source relationships, editorial judgment, and accountability. AI handles tedious research tasks so journalists can focus on those irreplaceable skills.

How do I convince my editor to let me use AI?

Frame AI as a research accelerator, not a writing shortcut. Show specific time savings on tasks like background research, data analysis, and document review. Emphasize that you verify everything AI produces. Propose a trial period with clear guidelines.

Should I disclose AI use to readers?

Follow your organization's transparency policy. Many newsrooms now include AI disclosure statements. At minimum, never imply that AI-generated research is original reporting. The AI prompts for content creation guide covers disclosure best practices.

Ready to Level Up Your Prompts?

Stop struggling with AI outputs. Use SurePrompts to create professional, optimized prompts in under 60 seconds.

Try AI Prompt Generator