ChatGPT butchered my work

It happened again last week. A client came to me, slightly embarrassed, asking for advice. Someone had taken a press release we'd carefully crafted together and run it through ChatGPT. The result was a completely different format, tone, and structure that was really quite different to our original. The intention was good — they wanted to "improve" the copy — but the outcome was a great reminder that not everything that comes out of ChatGPT is fit for purpose.

The first time someone put my work through AI, I'll admit it bothered me. I remember thinking, 'AI has butchered my work'.

Reframing

I've since realised it's actually a good sign. It shows people are engaging with communications, and they're being proactive in their input and thinking about how to make it better. That's a win.

My advice in these situations is to firstly, not call out the use of AI at all. Instead, be the great communication consultants we are by explaining why information needs to be presented in a certain way, why tone and messaging matter, why specific keywords belong in particular places.

Then get to the crux of the feedback, what were they actually trying to achieve with the edit? You can usually tell.

This approach turns a potentially awkward conversation into a strategic discussion about communications effectiveness. Rather than making someone defensive, or even embarrassed, by their tool choice, you're keeping the conversation to the topic that matters — the work.

I am very pro AI and pro tools that elevate our work. There are moments where full delegation to AI is possible, and there are (many) moments when that is not appropriate. Anthropic's AI fluency course — which I highly recommend — highlights two critical skills: delegation and description. Both are essential when working with AI, and they reveal a deeper truth about our profession.

Yes, AI can help with initial drafts, brainstorming, formatting, summarising, proofing and the rest. But delegation requires understanding exactly what you're asking for and why. And description — explaining the nuances of audience, context, and strategic intent — requires human expertise that algorithms cannot yet replicate.

Using AI in PR

There is a danger in leaning too hard into AI. I recently read a story about a film animation agency who went bust after relying on AI. The quality of the work suffered, they'd lost their creative edge and clients noticed. It's a cautionary tale about the difference between using AI as a tool versus letting it replace human expertise.

As communications professionals, we understand context and feelings that we can sometimes not even articulate. We know that a single word choice can spark the right reaction - or the wrong one. I keep seeing LinkedIn posts with lists of words or moments that gives us 'the ick' — it's not always logical!

We also understand industry dynamics, stakeholder sensitivities, and the subtle art of timing. We can read between the lines of a brief and anticipate unspoken concerns.

Most importantly, we understand that effective communication is about human connection. Yes, generative engine optimisation (GEO) is ramping up and creating communication for bots is now highly important. But, in a world where information is abundant and easily accessible, we — comms professionals — shine in crafting messages that convey more than just information, we strive to create understanding, build trust, and inspire action.

So, when someone puts your work through AI, resist the urge to take it personally. Instead, see it as a moment to better connect with them and understand their intentions.

There is a nuance in reputation shaping, a delicate balance of stakeholder communication, an art of saying just enough, but not too much. My advice is to keep showing up with that expertise, keep delivering the highest quality work, and above all, keep being human.

Our goal shouldn't be to compete with AI. It's here, and it brings a lot of opportunities for us all.