Big changes at a top PR platform (and they involve AI)
Here’s what PR pros need to know about the new AI detector at one of the most popular pitching platforms.
Joni Sweet is a seasoned writer, editor, and content strategist with 13+ years of media experience. In this newsletter, Joni reveals insider tips to help PR professionals sharpen their pitches, avoid rookie mistakes, land coverage for their clients, and build stronger relationships with writers. Book a Flash Consulting session with her to get journalist’s perspective on your biggest PR roadblocks.
Breaking news: Qwoted is taking a tougher stance on AI-generated content in media pitches and expert quote submissions. Until now, it’s been using GPTZero to flag when a pitch or quotes submitted to a journalist might be written by ChatGPT or another AI program. But today, the platform announced it’s scrapping that tool in favor of a new partnership with Pangram, which calls itself “the most accurate technology for AI text detection.”
(Psst… if you haven’t tried Qwoted yet, use code JONI to get $50 off your first month and two free pitch credits. It’s one of my favorite HARO alternatives!)
My take as a journalist
OK, at risk of alienating myself from nearly all of my colleagues who publicly rail against AI… I don’t actually care that much if a pitch, quote, or even a piece of content from a professional writer had some AI help. The only thing I really care about is whether something I’m reading is accurate, original, and interesting.
The problem is, many publications flat-out ban AI-generated content, which means if you slip me an AI-written quote and I include it in my article, I’m in violation of my contract, which could be a death sentence for my relationship with that editor and publication.
(I talked a bit more about this at the Story Seeds Pitch Clinic with Pitchcraft. So if you want the replay, sign up for Pitchcraft and use code JONI200 to get $200 off your first month or JONI400 to get off $400 your first annual membership.)
Honestly, I believe every journalist who wants to be sure that the quotes they gather are truly from a human should get on the phone or video chat and do a real interview—it’s what I try to do for most of my reporting. But given the low rates writers are paid, many of us (including myself at times!) accept written quotes to help us work more efficiently and hit our deadlines.
And that’s where it’s helpful that Qwoted has partnered with Pangram to check submissions for AI generated content. I’m pretty skeptical of AI detectors—some of them have even flagged my old drafts as AI-generated, even though they were written well before generative AI existed—so I ran a few of my writing samples through Pangram to see how it did. I was impressed with how accurate it was. It’ll definitely make me raise an eyebrow next time the integrated tool flags a submission on my Qwoted dashboard as AI-generated in a way other detectors never did.