Our AI Policy
By Christy Harrison, MPH, RD
This policy is intended to outline our approach to using—or mostly *not* using—AI here at Rethinking Wellness.
As the sole author of every written piece on this platform (so far), here’s how I use generative AI in my writing:
I don’t.
I don’t use it to think up ideas for me.
I don’t use it as a “thought partner” to bounce my own ideas around with.
I don’t use it to generate first drafts.
I don’t use it to edit or proofread the articles I write.
I don’t use it to research what the science says or find studies for me. (I do sometimes ask chatbots questions to see how accurate or inaccurate their answers are and write about the issues that arise.)
I don’t use generative AI in any other way for writing. Several years ago, I experimented with using a chatbot to polish my drafts and found that it stripped out my voice and made everything a little less accurate, while sounding “better” on the surface. It took me more time to go through the bot’s changes line by line (and usually revert them back to my original language) than it would’ve taken me to just write and edit the piece myself in the first place. I also tried using AI as a research assistant, but I found that it brings up too much misinformation, hallucinates false facts, references studies that don’t exist, and doesn’t find what I actually want it to find. I have my own system for research that I’ve developed over 23 years as a health/science journalist, and it may be a bit more time consuming, but it’s the only method I trust. Plus, it helps my writing process: I like to dig deep into the evidence and do the sometimes-boring slog of shaping quotes and ideas into a coherent piece, so that I can learn and make sense of the facts. If I were to use AI to summarize the research for me, it would bypass this kind of critical engagement.
When it comes to writing headlines, I use a headline analyzer that claims to be “leveraging AI,” though I was using it long before AI got involved and haven’t noticed a huge difference. I still come up with the original headline ideas myself and use the analyzer to improve them, weeding out anything too clickbait-y or inaccurate.
If I interview someone for a reported piece, I use AI to transcribe the interview and then manually edit any quotes for accuracy and readability.
If I ever publish guest posts on this site, I’ll hold them to the above standards.
I also interview guests for the Rethinking Wellness podcast, which appears here in both audio and written forms. Here’s how AI factors into podcast content:
My production team uses an AI transcription tool to create rough transcripts of the podcast, which a human editor then cleans up.
We don’t use AI voices to make changes to audio content. If any changes are needed, we cut or I re-record and my team inserts the new audio.
I will always interview real people and will never use AI to impersonate human beings.
If in the future I ever create an episode critiquing AI influencers (fake “people” created entirely by AI for the purpose of selling products—yes, this is a thing), I’ll clearly label any audio clips as coming from an AI.
My team and I use images to illustrate each post, mostly so that our homepage isn’t just a bunch of boring text. Here’s our stance on AI illustration:
We don’t use AI for image generation or editing. We believe in supporting real artists as much as possible, within the confines of our (shoestring) budget.
We pay for stock images from a platform that doesn’t accept AI-generated photos or illustrations. I hope that’s enough of a guardrail to keep AI visuals off the site, but if we ever discover that an image we used was secretly created by AI, we’ll replace it.
This is a working document and will be updated as technology and my work evolve. I don’t plan to incorporate AI in any other ways going forward, for the reasons outlined here—but if that ever changes, I’ll let you know. For now and the foreseeable future, Rethinking Wellness remains proudly human-made.
This policy is current as of June 12, 2026.