Photo credit: Zulfugar Karimov via Unsplash
By Press Kitchen VP Amanda Drum
If you worked in PR in 2024 or started work in PR in 2025, curiosity may have gotten the better of you at least once: can ChatGPT write a press release? Can Gemini craft a pitch to a journalist? Will a tool named Grok take my job in 1-5 years? (No offense, Grok.)
My short answer, if you want to stop reading here, is: no. At least, not yet. Too much of PR is predicated on sharing information that no one but your team knows yet until you send it to the media. For now, generative AI can’t predict the future; when that day comes, we’ll have bigger fish to fry.
That said, we can’t pretend it doesn’t have some impact on our writing roles today. We may not be using AI to draft releases or pitches, but other individuals use AI to write lots of different things: newsletters, blog posts, essays, bad cold emails. Because it has varied experience and is a machine-learning program, it adopts certain habits that, when an actual human replicates them, can make even 100% human-born copy look suspiciously “like AI”.
Some individuals began accusing online creations of being AI-generated, starting with digital art and animations, where they analyzed every digital brushstroke for “tells” that, unfortunately, were common traits of human-made renders as well. The habit has bled to the written word for the same reason: AI borrowed some grammatical habits. ChatGPT stole my beloved em dash.
Which brings me to where AI and PR intersect: how publicists can avoid making their writing look AI-generated, lest pitches be glossed over completely.
- RIP the em dash. You know, the long one.
I started my writing journey as a chronic semicolon abuser. I still am, but at least I diversified it with the em dash: a long dash that splices the middle of a sentence with an interjection that provides important context. Sadly, generative AI tools also love em dashes, and use them liberally…and sometimes incorrectly. To make matters worse, some devices make it challenging to create an em dash on a keyboard without knowing the specific shortcut; a point many accusers make if they see text overloaded with the punctuation. I have started editing them out of my team’s work, sadly, and will probably do so for the foreseeable future, in favor of the double dash: –, or the less fun en dash-plus-spaces: – .
- It’s not just you—it’s also you.
This one took some pointing out, but when you see it once, you can’t unsee it again. For some reason, AI writing tools love the following sentence structure: “We’re not just X—we’re also Y.” “It’s not only X—it’s Y.” Note the dash.
Honestly? I’m not sure why it loves this pattern so much. Maybe because it’s a common em dash application? What makes this discovery more insidious is that it’s not uncommon at all in regular writing. Humans make remarks like this all the time! I added this to my editing repertoire, too, in favor of positive upturns like: “We’re X and also Y,” or “It’s X and Y.”
- 👋
This may only apply to a small few of you in the social media marketing realm, unless you’re naturally ✨whimsical ✨, but sometime in the last year, programs discovered emojis and fell head over heels in love. Need a bulleted list? No, you don’t, you need a bunch of 👉❗💡.
I love emojis. Ask my team and check my Slack messages. What hits better to you: “Thank you,” or “Thank you! 🎉”? I also refuse to edit this particular quirk just because some ones and zeros also like a well-placed 🏆, unlike the other two items on this list. I just pick different symbols, depending on whether I’m talking to Gen Z-ers or fellow Millennials.
Thank you for allowing me space to vent. If these are my biggest gripes, PR remains in a fairly safe space for now. Still, it’s important to follow updates in AI for sophistication, new modes of learning, and anything else that could make it more stealthy–and thus more likely to overstep into publicity territory–in the future.
##