Cathy Resmer: Introducing Seven Days’ Public AI Policy

I have a confession to make. While compiling the Seven Days Vermont Cannabiz Guide in April, I used artificial intelligence to write a headline. The story, by Ken Picard, details ways that local cannabis dispensaries cultivate loyalty among their customers.

I edited the piece on a Friday afternoon after a long week and was struggling to come up with a creative title. After 15 minutes or so, in desperation, I did something I’d never done before: I turned to Claude, my AI chatbot of choice, asking it for 10 two- or three-word headlines for the story. “Looking for clever wordplay,” I added, wondering if it would be up to the job. No. 2 on its list: “High Fidelity.”

A great headline, but not my proudest moment as an editor. Seven Days has had an internal AI policy for more than a year now: We don’t allow writers to use AI to write articles. But headlines are packaging elements supplied by the editors to entice readers. At the time, they fell in a gray area. To brainstorm witty headlines, Seven Days editors routinely google idioms and pop culture references or use a thesaurus.

This was different: I asked Claude to do my thinking for me. I could feel my own wordsmithing skills shriveling ever so slightly. Conflicted over whether to use the headline, I consulted our chief wordplay whiz, culture coeditor Dan Bolles, who writes the short article teasers that appear on the cover of Seven Days.

Dan loved the headline but shared my reservations. In the end, we decided that using AI in this way is lazy and we shouldn’t do it — again, that is. We couldn’t resist running “High Fidelity” but changed our policy to prohibit that use in the future.

I’m in good company: In the latest issue of Wired magazine, global editorial director Katie Drummond describes her own strikingly similar dilemma. She came to the same conclusion Dan and I did, though she didn’t mention codifiying it.

AI companies and their tools are inescapable.

It seems like everyone is debating AI right now. From business to politics to entertainment to health care, AI companies and their tools are inescapable. There’s a story about how Vermont educators are using AI in the Back to School Issue of Kids VT, inside this week’s paper.

The technology is advancing faster than our ability to regulate it or even come up with best practices for using it. What can we do with AI? What should we do with it? Is avoiding it even possible?

To be transparent about how Seven Days is meeting this moment, and to be accountable to our readers, we’ve made our AI policy public. You can find it at sevendaysvt.com/AI-policy.

It affirms that Seven Days stories are written by humans; that the photographs of people, places and events accompanying our content were not created by AI; that when it comes to illustrating our work, we prioritize hiring local artists.

It also covers allowable uses. Writers routinely rely on AI to transcribe interviews — a huge time-saver — and to analyze data. AI tools can sift through thousands of documents quickly and have the potential to supercharge our accountability journalism.

Though we’re finding ways to incorporate AI tools under the hood of the car, humans will always be in the driver’s seat. And when it comes to our journalism, you can count on us to be up front with you about how we’re getting from point A to point B.

I’ve learned my lesson when it comes to headlines. However, I was excited this summer when Seven Days intern Olivia Fleming used AI to build a leaderboard for the Good Citizen Challenge, our youth civics project, to keep track of participants’ progress. She described what she wanted to Claude, and it wrote the code, a process known as “vibe coding.”

That was Collins Dictionaries’ 2025 word of the year. Like it or not, I suspect you’ll be hearing more of it in the future.