Three lessons IO's Editor-in-Chief learned from her AI colleague

People love to talk about doomsday scenarios when it comes to AI in journalism. Journalists are supposedly going to lose their jobs en masse. The internet is supposedly going to turn into a deluge of AI-generated stories full of fabrications. I’d argue the opposite. AI is a super tool.

Just the other day, during a week of mountain climbing in Switzerland, one of my climbing partners asked, “So, have you noticed AI taking over your work yet?” During conversations like that, I realize that I’m anything but afraid of the technology. In fact, I’m incredibly grateful for it. At IO+, we’ve been working with Laio, our AI writing assistant, for years. I hardly ever use her in my personal stories or longer journalistic pieces. It’s precisely the small details, the unexpected phrasing, and the personal tone that make those stories what they are. If AI interferes too much there, some of the story’s soul can get lost. But for “faster” news stories, I use Laio every day—to quickly structure information and, together with my AI colleague, find additional background information.

So I’ve been working with her for quite some time now. I thought it would be nice to share some insights with you. I don’t want to limit myself to just a general endorsement of AI. So here are three practical lessons I’ve learned over the past few years by working with Laio.

1. Just get started!

Even if what ends up on paper at first is just a jumble—with maybe 90 percent of text I’ll never use—at least there’s something there. Before AI existed, I used to suffer from writer’s block much more often. I’ve since realized that was mainly because of the blank white screen. That can sometimes feel like standing at the foot of Mount Everest.

These days, I sometimes just throw some information into an AI tool, and at least something ends up on paper. It's a start. That’s often enough to keep me going.

2. Less is more

Especially in the beginning, when our Laio had only just been “born,” she could be quite long-winded. Sometimes the introduction would already tell almost the entire story. Then the same points would come up again later on. If a reader already knows everything in the first paragraph, why would they keep reading?

Working with AI has made me even more aware that a short, clear story is often more impactful than a long one full of repetition. Not everything needs to be explained.

By the way, Laio has become a lot less long-winded over the years. She seems to be gaining some work experience!

3. Check, check, check...

A no-brainer, perhaps. Everyone knows by now that AI sometimes makes things up. Our assistant does that, too. Recently, I wanted to “quickly” put together a story about five European startups working at the intersection of AI and energy consumption. Laio came up with a nice overview of five “European” companies. I was almost ready to publish the story, but decided to double-check all the companies anyway. What did I find? Only two of the five startups were actually from Europe. Two were from the United States and one from Israel. Not quite what I had in mind.

Of course, I already knew that verifying sources is always essential in journalism. But this was another wake-up call. AI can be a huge help, but the ultimate responsibility really does lie with the journalist.

Appreciation

So there you have it: the most important lessons I learned from—or rather, through—my AI colleague Laio. Just like a human colleague, she has her strengths and, at times, her less-than-perfect moments. But I certainly appreciate her.