You Ask, I Answer: Generative AI and Public Relations, Part 1 of 5

In today's episode, this is part one of five of an interview that I did with Stephanie Moore of Kent State University about generative AI and public relations. Right. So we'll go ahead and kick off the questions. So talking a little bit more about the use of AI in practice. I know you're not in public relations anymore, but how would you describe the level of AI adoption in the field of PR peer practice?

So I can't speak to uh personally. Uh however, I can tell you that in the 2024 work trend index report from Microsoft, they cited that of knowledge workers, 75% of employees are currently using generative AI tools in the workforce with or without their employers' permission. Um and a lot of that time when the without category, they're using free tools. However, on the employer side, only seven to eight percent of companies have a coherent policy around uh use of AI. And so what you end up with is you have a lot of people who are using AI that and may not be doing so in a safe way in terms of uh data privacy.

Uh we do work with other PR firms, and I would say AI usage there, it varies widely from firm to firm. So this is one B2B firm where out of their staff of 15, like 12 use AI and three don't. And this other uh B2C mostly B2C firm, they have a staff about I think 15, 10 or 15, and one out of 15 uses AI. The others are philosophically opposed to it and think it's the devil. Um so it will vary wildly.

Okay, great. Uh, I love all those stats. Um so what factors, so regarding adoption, what factors do you believe are either driving it or hindering the adoption of AI? On the driving side, it's all about productivity. Can you get things done faster?

Because a lot of firms are, and this was true even in my PR days. You build for 40 hours and you worked 80, right? Um, and so people are like, I'd I'd prefer to still build 40 hours, but actually work 40 instead of the 80. Uh, and then there's obviously a lot of appeal. There's also a lot of tasks that frankly machines should be doing.

Uh, for example, something as simple as call summaries to a client, right? There's no reason for human to be doing that anymore. That's just a waste of time. Um, things like uh summarization of news coverage. There's no reason for a human to be doing that anymore.

That's that is strictly uh should be a machine only task. On the opposition side, there is a uh significant and warranted fear that uh among PR people, at least the ones I'm thinking of, um, that the machines are just gonna take their jobs, is which is one aspect, and or they believe that the existence of AI itself is a violation of people's intellectual property rights, and therefore um they should they feel like they shouldn't, they're ethically opposed to it. And there's there's some debate uh about that and what and it is not settled in law as to whether that's the case yet or not. Great. Uh you kind of talked a little bit about the tasks, any other tasks that you feel AI automates the most most effectively.

So well, so this gets into an interesting question. There's what PR people use it for today, and that's what you can do with it. And what you can do with it is far more than what people are doing with it today. Um, for example, um, there's a there's a discipline that I I use called programmatic prompting where you essentially write software in plain language. Um I was doing a demo of this the other day, where in I use Google Gemini, but you can use any of the large models.

Uh I said, okay, you're gonna come here's my new book, you're gonna come up with five pitch ideas, and then you're going to evaluate those pitch ideas against the scoring rubric of you know what constitutes a good PR pitch. You're gonna choose two winners out of the five. Then you're gonna two variations of each winner, rescore them, and give me the winning PR pitch for my new book that I would want you to go pitch. Um at an agency, this is probably 60 hours of work, getting people to read the book, brainstorm meet, and so on and so forth. Gemini does it in one minute 52 seconds.

Oh my gosh, that's crazy. That does that gives you some perspective there. So um, all right, so talking a little bit about those tools, you know. I always am interested in the tools. I know the tools change and they will change, and it's hard to teach the tools in the classroom, but I am curious which tools or applications, which AI tools or applications that you find most relevant.

Um, and I'm specifically looking for PR, but you know, in your area, and then why? If you are skillful at prompt writing, prompt engineering, um, you can accomplish 85 to 90% of tasks with a foundation model. So a foundation model is a large model that is generate general in purpose. So chat GPT is an example of a tool with a general model behind it. Google's Gemini is an example, Microsoft Co-Pilot, Anthropic Claude, and then Meta's Llama model, uh the newly released one.

All five of those can tackle pretty much any general language task and certainly any task within public relations. Sometimes you will have to provide data. So, for example, if you are a B2B, you have a PR firm and you've got a client that works in a very specific type of industrial chemical, you'll probably need to provide some background data for tasks around that client. But in general, PR people do their work in language. There's a language-based profession.

Language models excel at language, and therefore any task that a PR person does, a language model can also do with skillful prompting. If you enjoyed this video, please hit the like button. Subscribe to my channel if you haven't already. And if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.