What to know about using AI in journalism

• AI works best as a support tool, not a writer. The most effective uses in journalism are editing, data analysis, and idea refinement — not generating full stories.
• Simple AI workflows can immediately improve speed and clarity. Tasks like AP style checks, extracting key takeaways, and analyzing documents help journalists work faster and produce more structured, scannable content.
• Verification is non-negotiable when using AI. Journalists should always question AI outputs, ask for sources, and treat it like an “unreliable assistant” that needs oversight.

There are easy, practical ways to bring AI into your journalism workflow without crossing ethical lines or coming across as cringey.

But first, the biggest caveat of all: What is your newsroom’s AI policy?

If the answer is “we don’t really have one,” that needs to move up the priority list. Fast. But, otherwise, anything I suggest here should be trumped by the counsel of your newsroom managers. Just so we’re clear.

That said, ethical AI use in journalism does not have to mean letting a chatbot write your stories. We like the writing part of our work! However, I use AI mostly to support what I’m already doing — reporting, editing, organizing, pressure-testing (second opinions), and thinking.

Used that way, AI can be genuinely helpful.

Here are four easy ways to implement it right now.

1. Use AI to check AP style and spelling

This is one of the simplest and safest uses.

Maybe you do not want an AI tool reviewing an entire story. Fine. You can still use it to answer one very specific question:

• Is this AP style?
• Is this term capitalized correctly?
• Am I handling this title the right way?
• Does this sentence have a grammar problem?

Used carefully, AI can act as a quick style and spelling backstop. This is not a replacement for judgment. This is not a replacement for editing. This is a support tool.

2. Use AI to identify three takeaways for AEO

AI can also help you spot the most useful, answer-friendly points in your story.

I’m not saying it has to write them. In fact, I would rather it didn’t. But it can be very useful in identifying the three clearest takeaways from a piece of reporting.

Why does that matter?

Because more journalism now needs to be structured for AEO — answer engine optimization. People are increasingly finding information through AI summaries, voice assistants, search snapshots, and question-based queries. That means your stories should make their core value obvious, up top.

One simple tactic: ask AI to identify the three biggest takeaways in your article, then write or refine those yourself and place them near the top of the story.

You can brand them however you want:

• Key points
• What to know
• The big takeaways
• Why it matters

Whatever fits your publication.

The point is to make your journalism easier to scan, easier to understand, and easier to surface in answer-driven environments.

3. Use AI to mine data and documents

This is one of the most useful applications in journalism.

AI can help you:

• organize messy information
• sort through datasets
• pull recurring themes from long documents
• scan large PDFs for specific terms or references
• identify patterns worth investigating further

Say you have a 50-page PDF and need every mention of a certain agency, person, or phrase. AI can help with that.

Say you have a spreadsheet full of public records and need to group, sort, or summarize it so you can find the real story. AI can help with that too.

This is where AI starts to feel less like a gimmick and more like an actual reporting assistant.

Not because it is doing the journalism for you.

Because it is helping you get to journalism faster.

4. Use AI as a second opinion — or your sharpest critic

This might be my favorite use of all.

AI can be a very useful second set of eyes when you use it as a critic instead of a creator.

For example:

• Here are the questions I plan to ask the mayor. What am I missing?
• Here is this story draft. Where are the weak points?
• Is there anything here that could create legal risk for the newsroom?
• What follow-up questions should a skeptical editor ask?
• What important context may be missing from this framing?

That is a powerful use case.

But here’s the other side of this: you need to be AI’s critic, too

If you use AI in journalism, you cannot be passive.

You have to challenge it.

You have to interrogate it. Do not just accept its responses!

You have to treat it with the same skepticism you would apply to any questionable source, rushed claim, or incomplete piece of information.

Ask it:

• How do you know that?
• Where did you find this?
• Show me the source.
• What are you assuming here?
• What might you be getting wrong?

Always double-check. Always verify.

AI can be a useful assistant. It is not an editor. It is not a lawyer. It is not a reporter. And it is definitely not above scrutiny.

Did I use AI to write this post?

All my posts start like this, in Evernote:

I take that script and use it for a vertical video recording, pasting the text into the BigVu app — a teleprompter for my iPhone.

But I also put it into ChatGPT and asked it to rewrite it as a blog post in a form that supports SEO and AEO. For example, see the three takeaways at the top of the article? That’s an AEO tactic. The structure of the story with the <H2> tags? That’s an SEO tactic. All AEO and GEO start with strong SEO. Anybody who tells you that SEO is “dead” is simply wrong.

As for the story’s cover image? I use Canva. Specifically, I look for YouTube thumbnail templates and customize them for this.

If this was helpful and you’d like to discuss AI policies or workflows for your newsroom or business, reach out to me here.