There’s a lot of talk about the AI bubble right now, driven in part by the perception that productivity gains from AI are largely illusory. I can’t speak to the market or whether AI is widely overvalued or undervalued, but I can say that in the past year alone, AI has completely changed the way I work.
If you look at the tools we have today that didn’t exist a year ago (deep dives, browser agents, significant performance improvements in all modern models, etc.), there are many ways that AI can speed up or enhance many tasks for knowledge workers, especially journalists. As an independent journalist, I’m probably a little less constrained than others (my AI policy is what I want), and AI has made me rethink how every part of my job is done.
Bubble or not, that change is already here. To explain this more clearly and highlight tools and techniques that are broadly useful, we’ve detailed several ways to use AI in writing, research, and reporting.
The most intuitive way to organize this is to follow the process of your story, from idea development to publication and sharing on social media. Here are 10 ways I leverage AI as a journalist and content creator.
- Beat monitoring: Each reporter has a different set of stories to follow, and this is where AI’s ability to personalize and summarize can be very helpful. I think ChatGPT Pulse is very useful for not only editing stories, but also explaining why the story is relevant to your work and including other personal information (schedule, etc.), but it’s limited to the Pro plan ($200/month). A stripped-down, but still useful tool is ChatGPT or Perplexity’s scheduled tasks. This tool creates prompts that find the latest articles you’re interested in and send them to your inbox.
- Accelerated skimming: When I find an article I’m interested in, I often read the summary before deciding whether to take the time to read it. Perplexity’s Comet browser has a dedicated button for this, and there are many browser extensions for Chrome that do this. I noticed that in X there is a feature that allows you to easily reply to posts with “@grok what is this post about?” A short, instant explanation can save you a lot of time, especially when it comes to memes and trends you’re seeing for the first time.
- Going deeper: When Google’s NotebookLM made a splash last fall with its ability to create instant conversational podcasts, many dismissed it as a gimmick. However, I find this feature to be a very useful tool for preparing topics and news articles. Drop a single URL into a folder or use the new Fast Research feature to search for stories, generate audio summaries, and boom! We create short podcasts about whatever we’re researching. You can listen even faster by listening at 1.5x speed.
- Story ideas: Finding connections between stories and missing angles is always a mostly human-driven process, and NotebookLM is a great partner here too. Encouraging them to suggest story ideas based on questions that are implied or unanswered by the series of articles in your notebook is often a great starting point for thinking about story ideas.
- Get into the weeds: Once you know the idea you’re trying to implement, it’s time to start researching. AI is clearly better, especially the in-depth investigation tools that have emerged over the past year. I use all of these, but in different ways. Perplexity is a great first pass. It’s fast, but usually not as thorough as others. In general, Geminis are good at finding sources on the web, but I don’t think they’re very good at prioritizing. My go-to for serious research is ChatGPT, which I think is the most thorough. It is also excellent for demonstrating deep research abilities. inward– Direct users to large sets of files in Google Drive or Dropbox folders (via connectors).
- Target very specific information: When the key to your story is a single document or piece of data, it’s useful to outsource the task of finding it to a browser agent. For example, court documents are typically stored in difficult-to-access services such as PACER (Public Access to Court Electronic Records). Just finding the right case often requires performing multiple searches, especially if you don’t know the exact names or specific courts involved. Given such a clear task, browser agents like Comet and ChatGPT Atlas are perfect research interns, typically finding exactly what you’re looking for in just a few minutes.
- Writing coach: Now that you’ve gathered and processed all your background material, it’s time to write. In my columns and independent reporting, I don’t rely on AI to write for me, but I often use it as a coach. I created a custom GPT to act as a curious interviewer. Ask some questions orally, record long answers, and then compile them all into an outline when you’re done. From there, you’ll be off to the races and maybe back to the coach if things go in an unexpected direction. This process allows me to write in about half the time I used to.
- Writing Intern: I write a news digest for Thursday’s newsletter. Each item is co-authored with AI. For these short blurbs, I built a Claude project trained to my style, digest format, and target audience. Given a story or news trend, it creates a one-paragraph summary in my style that I edit and add to. A 15 minute process takes approximately 5 minutes. (And yes, I’m completely transparent with the audience if the AI acts as a co-author.)
- Copy desk: Everything I write goes through the AI copy desk. First, Custom GPT will critically and constructively review it and suggest edits (and possibly further research) to enhance your work. Another GPT then runs a proofing pass, aiming to make it more newsy or conversational depending on the context. Importantly, neither GPT does automatic rewriting. Those are all suggestions I can take or leave.
- Social media manager: Finally, once your work is published, you’ll receive a specific prompt to create social copy on LinkedIn, both on your personal profile and on your Media Copilot company page. In this case, instead of manually pasting anything into Assistant on the web, use Zapier, an automation tool that automatically generates social posts and queues them for review. This saves you the hassle of switching between browser windows and ensures you never miss a post.
I hope you can understand the pattern I have explained. Each process involves the AI paying attention to things, but the AI prioritizes. my attention. In my view, AI greatly speeds up certain tasks in story creation, but the process still requires human review and judgment throughout. The AI can serve as a writing “intern” for very specific formats, but I leave the text alone for anything substantive (like this column) written by humans.
No matter how intelligent and automated parts of a process become, its human-centeredness remains. There’s no doubt that AI will continue to improve, but the purpose of using it is to sharpen, accelerate, and amplify the way you communicate with your audience.human audience. Machines can be great companions, but their focus comes when we stop communicating and simply create “content.” Media has just experienced an era dominated by SEO, algorithms, and least common denominator thinking. Let’s not do that again.
