40 Questions Journalistss Should Ask Before Trusting AI
AI can find information fast, but it cannot tell you why a story matters or who needs to hear it. These questions help you catch the moments when your AI tool is doing your thinking instead of helping you do it better.
These are suggestions. Use the ones that fit your situation.
1When Claude summarised that research for your background section, did you read the original sources yourself or just accept its version of what they said?
2If you used Perplexity to find expert names for a story, did you then ring those experts directly or quote from what the AI found about them online?
3How many sources have you actually interviewed for this story versus sources you know about only because an AI listed them?
4When you transcribed an interview with Otter.ai, did you listen to the full recording yourself or work only from the transcript?
5If a source gave you a complicated answer, did you go back and ask them to clarify it, or did you wait to see what ChatGPT made of it?
6Are there people in your story whose perspective you included because an AI suggested it rather than because you thought it mattered?
7When was the last time you built a story by ringing sources cold instead of starting with what you could find through AI search?
8Do you know the phone numbers of your regular sources by heart or do you search for them each time?
9Have you noticed yourself getting slower at recognising when a source is evasive or hiding something?
10If an AI tool is not available tomorrow, could you still report this story using only the relationships and instincts you have built?
News Judgement and Story Selection
11Did you decide this story matters because you recognise its importance for your readers, or because an AI tool ranked it high or an engagement algorithm suggested it?
12When Perplexity showed you trending topics, did you ask yourself whether those trends matter to your actual audience or just whether many people are searching for them?
13Can you explain why this story is worth publishing now instead of next month in language that has nothing to do with search volume or social media potential?
14If you used an AI to help choose between two possible stories, what would your own instinct have chosen and why?
15Are you chasing stories because they seem important or because they fit the angles that AI tools keep surfacing?
16When you read an AI summary of what a story is about, did you then think differently about which parts really matter?
17Have you published a story recently that was factually accurate but somehow missed the real point that people in that situation kept trying to tell you?
18How would you justify this story to an editor who did not have access to the same AI tools you used?
19Did you choose this angle because it answers what readers actually need to know or because it was easiest for an AI to help you research?
20When an AI suggested a different way to frame the story, did you test that frame against your own reporting or just use it because it was offered?
Fact Verification and Research Depth
21When ChatGPT confirmed that a fact was correct, did you independently verify it using your own sources or stop there?
22If Perplexity cited a source for something important in your story, did you find that source yourself and read the original context?
23Can you identify which facts in this story came from AI research versus facts you sourced directly and which ones you actually verified twice?
24When you used Otter.ai to transcribe an interview, did you notice errors in the transcript that might change the meaning of what was said?
25If an AI tool said a statistic was from a certain organisation, did you ring that organisation to check or trust what the AI reported?
26Are there any claims in your story that you would struggle to defend by pointing to a primary source you have personally reviewed?
27When multiple AI tools disagreed about a fact, which one did you trust and why, or did you do your own reporting?
28Have you noticed yourself accepting AI summaries of research papers instead of reading sections of the papers themselves?
29If you had to cut ties with all AI tools tomorrow, which facts in this story would you have to re-report from scratch?
30When Claude wrote your first draft, how many sentences did you leave unchanged because they seemed good enough rather than because you verified them?
Editorsial Instinct and Professional Judgment
31Before you asked an AI to help with this story, what was your gut instinct telling you about what the real problem was?
32Can you recognise the difference between a story that AI research made easier and a story that you would not have written without AI suggestions?
33When you were starting out in journalism, what questions did you ask sources that you no longer ask because you can search for the answer instead?
34Have you tested your story instincts recently without AI tools, or would doing so feel slower and less confident than it used to?
35If you showed this story to a reporter who does not use AI, could they identify any places where your reporting feels thinner than it might otherwise be?
36When you pitched this story to your editor, did you explain why it matters or did you rely on metrics and trending topics to make the case?
37Are there stories you decided not to pursue because they seemed hard to report rather than because they did not matter?
38Think of the best story you have written in the past two years. How much of the reporting for that story came from conversations with people versus research through tools?
39When a source tells you something surprising, do you still feel the instinct to dig deeper or do you wait to see whether an AI tool flags it as noteworthy?
40If you read back through stories you wrote three years ago, before you relied on AI, do you notice yourself making different choices about what to investigate?
How to use these questions
When you use Perplexity or Claude to summarise research, treat the summary as a starting point for your own reading, not as a replacement for it. Read the original source yourself before you rely on it in your story.
For every expert source that an AI tool suggests, make one cold call to a source that it did not suggest. This keeps your instinct for what matters sharp.
After you finish a story, write down three sentences explaining why it matters that have nothing to do with search volume, engagement, or trending topics. If you cannot do this clearly, the story may not be ready.
Listen to at least one full interview recording yourself before you rely on an Otter.ai transcript. Transcription tools miss tone, hesitation, and context that can change what a quote actually means.
Once a month, report a small story without using any AI research tool. You may be surprised at what your instincts tell you when you are not relying on a tool to find the angles.