Last week I was surprised by strong pushback on my beliefs about the impact of AI on PR. I spoke separately with two experienced marketers I respect very much. Even though they are tech-forward, both were pumping the brakes on my interest in incorporating AI writing tools into my speaking and training. Even though they don’t know each other, they both said that it’s too early to know if this is a trend or just a fad.
Because this caught me so off guard, I slowed way down and listened very carefully to their critiques. I thought a lot about it and researched it further. And I ended up back where I started: Applying AI writing tools to PR is absolutely a trend, not a fad.
That’s why I’m starting a new newsletter called “Michael Smart’s AI in PR Newsletter.” To help people distinguish between my two newsletters, this new one is on Substack – it will remain free of charge. If you’d like to receive it, please subscribe over there (and if you’d be so kind, share this link with other PR pros).
In early January I made specific predictions about the rate of adoption of AI tools for PR (republished on Substack here), and in February I responded to some early criticism of those predictions (republished on Substack here). Next are more recent criticisms, and my responses:
Right now it looks promising, but it’s still hype, and it remains to be seen what’s going to work
It’s not just because Microsoft is investing billions in this – Google too – that this is a valid trend. Lots of smart people in tech invested billions in crypto/web3 and . . . here we are. That analogy spotlights a big difference with AI writing: While web3 was all about the promise of decentralization, you and I are using AI writing tools right now. I used ChatGPT to write a better title for this article. If writing subject lines or headlines is the ONLY thing it ends up doing reliably well, that’s still a dramatic impact on all of our workflows. And it’s obvious it’s going to be effective at much more. In future editions of this newsletter, I’ll share case studies of junior PR pros using ChatGPT to finish in literally seconds what their teams thought would take 2-3 hours, and of PR leaders whose teams are already deep into using this tech to do their jobs.
That NYT columnist got the Bing chat to profess its love for him and encourage him to leave his wife
I’ll admit, reading the transcript of that chat is a weird emotional experience – freaky is not a freaky enough word. But I don’t work for the Bing chat team – I don’t need to tame that chatbot, and neither do you. All we need to do is apply it to what it’s already good at: writing first drafts of commodity content and saving us hours. We can then use those hours for more creative, strategic work.
People are using AI to write scams/instructions for destruction/horrible and offensive things
This is a legit concern that is paramount for the teams at the big companies who are building this technology. They MUST rein in abuse of these tools. I can’t guarantee it, but I believe they will. Bad actors use email and HTML every day for hurtful and offensive purposes and that doesn’t stop us from using those tools for good. The hard truth for critics is that the new AI tech is here to stay – even if the government stepped in to regulate it, or masses of Americans rose up and persuaded the tech companies to put the genie back in the bottle, overseas companies would continue to develop it, and work/business would flow to those companies. We as a society gotta figure this out, and I’m optimistic we will.
Hold up, ChatGPT isn’t actually that good at writing, professors can easily spot essays written with it
This is the most common “gotcha!” that I hear, and it’s either disingenuous or radically uninformed. Of course the writing is replete with factual errors and sounds like it was created from a template! The company behind ChatGPT, OpenAI, has been “open” about those weaknesses from the start. Anyone looking to use generative AI to totally replace humans deserves to fail. The reliable way to succeed with ChatGPT – or any of the interfaces built on top of it, such as Jasper – is to only apply it when you know more about your subject than your audience does. That way you can catch factual errors and delete the obvious. As of right now, I’m advocating this tech as a productivity tool, not a replacement for a human editor/creator. That said, I bet that all this real-life testing and feedback will lead to these tools getting better very quickly.
AI writing tools for PR are here to stay. I’ll be researching and experimenting with them for you – come along for the ride. Subscribe to “Michael Smart’s AI in PR Newsletter” on Substack. As always, tell me how you’re using AI in your work, or share with me your concerns about it that I haven’t addressed.
My original newsletter – “Win With Smart PR” – will continue to be posted here every Thursday and will continue to cover issues and tips most relevant to media relations professionals.
This article was originally published on March 9, 2023
(I’ll also send you other weekly tips)