Your clients are probably using AI to answer email questions. Here's why it's a bad idea
A look at what AI-generated interview responses actually look like so you can help screen your client responses for them
Happy Thursday Top Tier Consulting community!
This week we’re diving into a topic that has been on many peoples’ minds lately: AI.
We’ve touched on this in the past (Recommended reading: The Hidden Dangers of Relying on ChatGPT for Your PR Efforts) but wanted to dive in a bit deeper and talk about the latest ways in which we, as writers, are seeing this being used by sources. And it’s something that is a surefire way to make sure that A) Your source doesn’t make it into an article and B) A writer doesn’t ever want to work with that source again — even on a different topic for a different publication — AND it could even lead to C) A writer not wanting to work with a publicist agin (if the situation isn’t handled with care).
Writers are busy folks and are often balancing many articles at the same time. As a result of this sometimes we have to do email interviews (Recommended reading: How to Help Your Clients Rock Email Interviews). While some sources rock out their responses, following the tips that we recommend in the newsletter above, others are going down a dangerous road lately and using AI to craft email responses for them.
Before we dive in, a quick reminder that if you missed last week’s epic HGG workshop, you can buy the recording for the price of one dinner entree — surely, landing just ONE more holiday gift guide inclusion for your clients is worth that :)
We discussed so many crucial aspects of pitching for the holidays, including:
Examples of sample subject lines that were SO good, we opened those emails (and we receive 500-1,000 pitches a day)!
A checklist of everything you should include in your HGG pitches
Advice for how to make YOUR pitch stand out in a writer’s crowded inbox
Tips for getting your SOURCES included in holiday gift guides (great for clients who don’t have products!)
A discussion of what’s trending this year and what we’re seeing in terms of themes for holiday gift guides
Examples of HGG pitches we’ve received that hit the mark — use these as your inspiration for format, content, etc.
Subject lines to use that will help ensure your email gets opened (and which subject line emails get deleted right away)
Plus, we had a Q&A session, where we discussed:
Which affiliate platforms editors prefer these days
If PR should respond in Substacks, etc. or email writers directly
How to pitch multiple products in one pitch
If the election and 24/7 breaking news cycle is impacting holiday coverage
How important are product samples for gift guide consideration?
How to pitch children’s books, jewelry and other certain types of products
How to handle luxury products
How to cater a pitch to an individual writer vs. a mass pitching approach (and why this is essential)
Are the same writers who wrote them updating gift guides annually?
How to make a pitch look less spammy
And more…
Cost: $29 Paid TTC Substack Subscribers; $49 Free TTC Subscribers
(you can always upgrade for the discount AND it’ll give you access to our weekly paid posts and monthly Top Tier Talks, too)
Ready to buy it? Send an email to info@toptierconsulting.NET and we’ll collect payment and send you the recording.
Now back to this week’s topic.
Here’s a situation that happened this past week that inspired this newsletter:
Nicole recently put out a call for sources. A publicist wrote back and said that they had a client who would be perfect for an article and would like to do email interview questions. Nicole agreed to use them for a source, and they promised that the questions would be delivered by a certain date.
When Nicole received the answers, she was shocked to find that the source did not actually answer the questions, but basically wrote a paper that the themes of the questions into it. It was incredibly obvious to Nicole that this was written in AI. She has done some work for some corporate clients who have used AI and ChatGPT to write blog post and have been hired her to edit them to try to make them sound more human. So once a writer becomes a little bit familiar with how AI looks feels, it’s really obvious to us when something was written by ChatGPT.
What were some of the red flags that this was AI generated? First of all the source didn’t answer the questions. Nicole expected to receive reposed in a Q&A format and instead was given an entire research paper instead. So that was the obvious red flag. Second of all, the article had subheads, bolded sections, a title, a byline, footnotes within and used language that was very obviously AI generated, like “next. we will take a look at this.”
Nicole immediately emailed the publicist and asked, “did your source use ChatGPT to write this?” They said they didn’t think so. Nicole said, “please check with your source,” explaining to the publicist that writers are not able to use responses generated by AI. It’s against our contracts with many of our publications and we can get in big trouble for it.
The publicist came back and said that the source said, “oh I actually wrote it on my own, but I put it through ChatGPT to check for for grammar.” Hmm… This was obviously a line and it didn’t seem like the publicist believed it either. So the publicist wrote back and said, “I understand if you don’t want to use these responses given the circumstances” and apologized.
From the moment Nicole received the responses, it was obvious she wasn’t going to use them. But she followed up with the publicists to see if she was aware.
The bummer about all of this is that the source was very qualified to be in the story. Nicole gave him the option either doing a phone interview or to do an interview over email and he said that he was busy and wanted to do email. So the fact that this person decided to have a robot answer the prompts instead of lending their expertise got them thrown out of the article. And now Nicole will never use that person again as a source. She’ll also warn her colleagues. The publicist handled this situation well on her end (but it would have been nice if they could have helped to find a replacement source for the article) but in general, this situation is something to be aware of as it could definitely strain a relationship between a writer and a couple assist.
This is the third time that we have seen this happened this week. We’ve submitted some calls for sources on platforms like Qwoted and had answers come back there with AI generated responses too. Never a good idea guys!
To dive into this deeper, we wanted to give you a little taste of what it looks like when a source uses AI to generate email quotes.
Here are a few real life examples, which can hopefully give you an idea, as a publicist, what to be on the lookout for. NOTE: We also recommend using an AI checker like Grammarly to see if your source used AI but it’s also worth noting that there was a robot discussion on Czars recently with writers saying those are oftentimes not accurate and they have been accused of using AI when they didn’t. So scan through this example so you can see what you’re looking for.
Shoot, there’s a paywall and our AI interview examples live beyond it! Not a paying member? Here’s what you’re missing out on:
Full access to our weekly newsletters
Access to our extensive archive of paid posts on every PR topic under the sun
The ability to attend/participate in our monthly Zoom Top Tier Talks
Discounts on all our most popular group workshops (like our upcoming Pitch Perfecting Workshop)
*Note: Our consulting sessions, workshops, Zooms and Substack newsletters are strictly educational. Signing up for anything has no bearing on landing coverage in any of our outlets. Our role is to fine-tune your approach and tactics so that you can apply these learnings when pitching other journalists.