Generative artificial intelligence is all the rage right now. I recorded a podcast episode last week with The Embargoed and we talked about fast it has come onto the scene—and in full force. It is taken hold of not just our industry, but medicine, law, engineering, home decorating, architecture, and everything in between.
In fact, Boston Children’s Hospital just announced they are hiring a “prompt engineer to work on large language models such as ChatGPT so that it can identify use cases and train its workforce on the most appropriate uses for the emerging technology.”
This feels very much like the advent of social media, when everyone was trying to hire specialists with 10 years of experience, but it had only been around for six months. The difference is that many business leaders proclaimed social media was just for the kids and it would never work in a work setting and they buried their heads.
This time around, though, pretty much everyone has at least tested out one of the AI tools, if only to see what the fuss is about.
A recent Muck Rack survey shows that 28% of respondents already use it and 33% plan to explore it. When you look solely at the C-suite, those numbers shift slightly, with 38% saying they already use it and 32% saying they plan to use it.
They’re not sticking their heads in the sand this time. They’re fully embracing it.
There Is Nothing to Fear!
Late last year, I explored ChatGPT and thought it’d be fun to write a blog post using a script generated by the tool. I was surprised at how easy it was to use and how it created a compelling first draft.
It wasn’t the final draft, but I loved how it helped me past writer’s block on a topic. Fast forward six months later, and I find myself using it every day to help get unstuck on certain things.
Just the other day, a client presented a pretty complex topic and asked me to help him dumb it down. I read the abstract and had no idea where to start. So I dropped it into ChatGPT and prompted it to help me.
It spit out copy that explained it to me in layman’s terms and then I was able to write an article for the client. Without that, I would have spent hours on research before I could have started. Instead, it took me 15 minutes.
I know many of you out there fear this will take over your job, and you’ll be left sitting outside Starbucks with a cup to collect coins. That isn’t the case at all! If anything, it’s going to make you more efficient, so you can do one of two things: take on more work (if you’re a masochistic workaholic like me) or free up time to do other things you enjoy.
Instead of spending hours on research, you can get up to speed in just a few minutes and go about your normal job with time saved.
Generative AI Prompts
You can use generative AI in your role as a communicator in many ways. HubSpot just did the hard work of putting together 50 prompts to use. They broke it down into buckets: lead generation, social media, audio and video content, content promotion, and repurposing old content.
Some examples include:
- Repurpose a blog post into a video script using this article: [insert old blog post]
- Turn a webinar into a podcast episode using this pre-existing transcript: [insert old webinar transcript].
- Repurpose an ebook into a series of [number] blog posts using this pre-existing text: [insert old e-book content].
- Rewrite a blog post into a social media post series of [number] posts on [social media platform of choice].
- Write a guest post for a popular industry blog discussing the impact of [product] on [marketing strategy].
- Suggest the [number] best hashtags for a social media campaign on [social media platform] to reach [target audience].
I’ve tested some of these prompts, and they’re not without their faults. For instance, I took a 4,000-word white paper and asked it to repurpose it into five blog posts. It spit out 100-word descriptions for five blog posts, so I had to take one at a time to get what I really wanted—and, even still, it’s only a really good first draft.
Please Fact Check
Kashmir Hill, a tech reporter at The New York Times, recently posted that she got a “PR pitch that expressed admiration” for a book she did not write. In fact, said book does not even exist.
The PR pro said, “We have enjoyed your coverage in The New York Times, as well as your book, ‘The Secret Life: A Book About the Future of Privacy,’ and we believe your expertise and knowledge…” Blah, blah, blah.
Go ahead and Google the name of the book. It does not exist. So Hill replied to the PR pro and asked where they got that information. They admitted they had used ChatGPT to write the pitch. AND DID NOT FACT CHECK IT.
Listen, my friends. If you are going to use generative AI to make you more efficient at your job—and I think you absolutely should—fact check the work. What the AI spits out is a great first draft. It should never be your final draft (at least, not right now—that could change in the future).
Don’t send things willy-nilly without editing and fact checking. Those are the two things you must do every time.
Generative AI Guidelines
This leads me to the guidelines just released by The PR Council.
- Always protect company information. This means as you’re training generative AI to produce content in the voice of the organization, don’t give it any confidential information to consider. This includes business plans, prospect or sales documents, paid or confidential analyst reports that you would like summarized, paid market insights to detail key findings, confidential research data, or text related to sensitive internal employee communications.
- Be ethical. Don’t use AI to create deep fakes or mis or disinformation. Use it for good. It’s that simple.
- Fact check, fact check, fact check. I won’t beat this dead horse (am I still allowed to say that in 2023?) anymore, but please do not send anything that has not been edited or fact checked. Don’t be the PR pro who sent a pitch to the tech reporter at the Times without checking it first.
- Disclose if you are using AI. To me, this one isn’t as critical. Do I need to disclose that I used AI to write a first draft of an article that I drastically changed in the editing process? My thinking on this one is to disclose it if it played a substantial role in a finished piece. If you used it just for an outline, I don’t think it’s necessary.
- Remember diversity and inclusion. I am a big fan of Canva. We use it pretty regularly for content. And one of the things that drives me absolutely batty is that most of the images in their library are of white people. So much so that I’ve asked my team not to use images that include humans in any of our work—and we’re working on building a library of our own. AI is built by humans, so it has the same biases as ours. Be cognizant of that fact when using it to create content.
- Spin Sucks. Establish clear guidelines and conduct regular training on best practices and ethical and legal use of AI. It should focus on all of the things we’ve discussed here today.
The bottom line is AI is amazing. It will change how we do our jobs—for the better. It’s not something to fear, but something to embrace. With some of what we’ve discussed here today, I’m confident you’ll be well on your way to being an AI fan.
