<img height="1" width="1" src="https://www.facebook.com/tr?id=272494640759635&amp;ev=PageView &amp;noscript=1">

How AI is transforming strategic communications

By On Top of PR

On Top of PR podcast: AI strategy for business communications with Kathleen Perley and show host Jason Mudd episode graphic

Insights on the world of AI and human-centric ways to utilize it for business strategies.


Tune in to learn more!


Our guest

Our episode guest is Kathleen Perley, professor of AI at Rice University’s Jones School of Business. Kathleen is an innovative healthcare tech leader and trusted adviser in the emerging world of optimizing AI for business marketing and advertising.


Watch the episode here:


Listen to the episode here:

Watch the podcast on Youtube.   Listen to the podcast on Apple Podcasts button.      Listen to the podcast on Spotify button.      



5 things you’ll learn during the full episode:

  1. The race for AI: staying authentic while building public trust
  2. Unique ways to use AI for business and personal development
  3. Top recommended AI tools
  4. How to utilize AI safely: best practices
  5. AI and deep fakes: What is coming next?


  • “Do we let this ego race to be first end up hindering our ability to have a public that trusts and aligns with AI and doesn’t build this resistance?” - @Kathleen Perley
  • “This [AI] is our generation’s moon landing.” - @Kathleen Perley
  • “How do we take the mundane, the rote out of our lives so we can spend more time doing what we do best, which is developing these relationships and human connections?” - @Kathleen Perley
  • “It's one of those things I keep hearing over and over again — that if you use AI to help generate content, you’ll get devalued in the algorithms from a search engine ranking perspective — but I’m not seeing that come true. I think the focus is more on quality and less on who is the creator or origin. If we use both humans and AI, it usually produces something even better than just a human or just AI alone.” - @Kathleen Perley
  • “If you haven’t started integrating AI into your crisis communication plan, now would be a great time to start thinking about that.” - @Jason Mudd 
  • “My biggest fear right now is that [AI] produces so much content so quickly that it almost takes just as much work to go through and improve it and fix it. But that first draft is awfully nice to have most of the time.” - @Jason Mudd

If you enjoyed this episode, please take a moment to share it with a colleague or friend. You may also share your experience with others by leaving us a quick podcast review.

About Kathleen Perley

Recognized as one of MM+M’s Women to Watch 2024 and with her agency on the Inc 5000 list for three years, Kathleen is a celebrated leader in merging AI with business strategy, especially in marketing and advertising. 


She also serves on MD Anderson's advisory committee and supports Ncourage, an investment fund aiding women-owned businesses.



Additional Episode Resources:

Our On Top of PR sponsors

  • Production sponsor: Axia Public Relations, one of America’s Best PR Agencies, according to Forbes Magazine
  • Presenting sponsor: ReviewMaxer, the platform for monitoring, improving, and promoting online customer reviews
  • Coffee Sponsor: Fans like you fuel our efforts using buy me a coffee.

Enjoy the Podcast?

If you did, be sure to subscribe and share it with your friends!


Post a review and share it!

If you enjoyed tuning in, leave us a review. You can also share this with your friends and family. This episode can give you professional insight into media coverage. Know your rights and the regulations to follow when it comes to the media.


Have any questions?

You can connect with us on Facebook, Twitter, and LinkedIn.


Thank you for tuning in! For more updates, visit our On Top of PR website or join the community. You may also tune in on Apple Podcasts, Spotify, and YouTube.



00:00:00:00 - 00:00:09:29


Welcome to On Top of PR with Jason Mudd, presented by ReviewMaxer.


00:00:09:29 - 00:00:24:06


Hello and welcome to On Top of PR. I’m your host, Jason Mudd, with Axia Public Relations. We've taken a quick break from On Top of PR, but we're back in full force and excited to be here. Glad that you're here, and I know you're going to be glad to be here, too, because today we're joined by my friend Kathleen Perley.


00:00:24:08 - 00:00:53:29


Kathleen is currently a professor of AI, that's artificial intelligence, at Rice University's Jones School of Business. She's a celebrated leader in emerging and merging AI with business strategy, especially in marketing and advertising. She previously owned an advertising digital agency that she grew to $17 million in annual revenue with more than 70 employees before it was acquired by private equity.


00:00:53:29 - 00:01:29:13


So again, congratulations on that. And Kathleen, she is, she and I met, we're both active members in the Agency Management Institute, and Kathleen uses her expertise to help advertising, public relations, marketing, communications, executives, and agency owners integrate advanced technology into their operations. Outside of the office, Kathleen balances her professional accomplishments with a vibrant personal life as a mother to Charlotte and Everett and enjoys refurbishing her 1972 International Scout.


00:01:29:13 - 00:01:57:02


I did not know that. And Kathleen is one of the smartest people that you can meet in the business. She was recognized as one of MM+M’s Women to Watch in 2024. And with her agency on the Inc 5000 list for three years, she serves on MD Anderson's advisory committee and supports Encourage, an investment fund aiding women-owned businesses.


00:01:57:02 - 00:02:17:13


Good for you. I love that. Her commitment to innovation spans her professional endeavors and community. Kathleen, that's a mouthful, but we're so glad that you're here. I'm honored to have you be another great guest on On Top of PR and excited to talk artificial intelligence with you today.


00:02:17:16 - 00:02:33:05


Yeah. Thank you so much for having me. I think, you know, anytime I can get somebody to nerd out with me on AI, my husband and all of my friends very much appreciate it, because then I'm not trying to ruin Friday-night date night or game nights with our friends talking about AI and nerdy stuff.


00:02:33:07 - 00:03:01:14


I love that, I love that. Well, Kathleen, we are recording this on May 29th of 2024, and I think that's important to establish that since, you know, AI is the Wild West right now and currently, you know, a new day, new issues, new opportunities, new tools, new features, new benefits, and so what we're talking about today, we are going to do our best to keep it evergreen for many years to come because the foundation and fundamentals are not going to change.


00:03:01:22 - 00:03:24:11


But the technology sure is. And that's just like PR, marketing, advertising, you know, the the core skill sets are the same. Just the tools give us more power, and with great power comes more responsibility. And so that's what we want to talk today about — is the responsible use of AI. And so let me just kind of tee it up to you, Kathleen, and kind of, again, late May, almost June here in 2024.


00:03:24:17 - 00:03:28:15


What's got you excited about AI in the last few weeks?


00:03:28:17 - 00:03:46:29


I think, you know, we're seeing a lot of changes in innovation. Obviously, we all are waiting on, you know, every second to see when ChatGPT-5 will be rolled out, but we're seeing a lot of progress. And, you know, I think about, right, I have a background in marketing, advertising. We specialize in the healthcare and wellness industries.


00:03:46:29 - 00:04:09:13


And so, you know, really do see, and we're starting to see some of those impacts and where, you know, this new technology could help us make huge breakthroughs, whether it's through — actually, there's a company out of Tulane University called Informatica who is doing some great work on utilizing machine learning to help identify bacterial resistance and help us kind of really fight those diseases


00:04:09:13 - 00:04:37:24


more from a strategic approach, which could ultimately save us thousands and thousands of lives, too. I don't know if you've recently seen they're utilizing AI and machine learning technology to help school systems really identify better bus routes because there is a shortage of bus drivers. And so, you know, while there is a lot of, you know, concerns around safety and security as it pertains to AI, and rightfully so, to a certain extent, there's so much opportunity and promise that's coming along with it.


00:04:37:24 - 00:05:03:11


And, you know, I even — there's a story recently, Google is doing this Project Green Light where they're actually utilizing AI and machine learning technology to synchronize lights so you don't have to wait in traffic as much, and so they're using existing infrastructure. There's not a lot of high cost to it, but they were able to reduce, in one of the cities, their traffic-stopping by 30%, which ultimately reduced carbon emissions by 10% in that city.


00:05:03:11 - 00:05:24:01


And so, just I think a lot of promise in terms of really being a part of the next breakthrough, you know, I think I hear my dad and his friends talk about the moon landing as, like, such an impactful moment. And I feel like this is our generation's moon landing. And so it's really exciting to be kind of in the trenches of it and kind of see it unfold.


00:05:24:03 - 00:05:44:22


Yeah, I completely agree. Somebody asked me the other day, I was on a panel and they said, you know, ‘What do you think people would point to as being, you know, the most interesting thing to have happened or changed in the public relations profession in the last, you know, 50 years?’ I think they said. And I said, ‘Well, I only know what's happened in the last 25, but I'm pretty sure it's going to be and is AI.’


00:05:44:25 - 00:06:04:14


And, you know, some other people were saying the internet and all that. And I just said, ‘Yeah, to me, if 50 years ago you were to show somebody the practice of PR today, right, they would be blown away by the internet, of course. But AI just takes it to a whole nother level because obviously you would need the internet to have the AI features and capabilities you do now.’


00:06:04:16 - 00:06:25:01


Yeah. And I think, you know, two things on that. One, the internet revolution happened over pretty decent length of term, right? It didn't happen as suddenly as we're seeing the AI revolution happen. And the other interesting part about that is, you know, the ability to and importance of PR, right? We talked about the impact, I don’t know, and I understand,


00:06:25:01 - 00:07:02:25


right, Google's AI search overview has had a big flop recently in terms of its performance. But if you utilize Perplexity AI or others, we're starting to see individuals move away from search engines and move closer to using AI as a way to gather education or information. And so I think all those years where, you know, in the digital, we were so focused on digital and so focused on building reputation through, you know, whether it's Knowledge Graph and showing up on page one, I think those are going to be kind of days past, and we're going to really have to focus on how do we garner true earned media?


00:07:02:25 - 00:07:20:10


And that's going to be a huge part of it. And PR is going to be leading the charge of that. So I think it's interesting. Brand and PR, you know, we've seen this big pendulum, right? Everyone, when I started my career, no one did digital. I was lucky if 30% of their budget was targeted to digital. Now I'm seeing companies where it's about 95% digital.


00:07:20:12 - 00:07:33:29


And there's that sweet spot. And I think, you know, the AI revolution, believe it or not, is going to push back for us to garner, you know, public attention through other channels and more traditional or more earned channels than we've seen in the past.


00:07:34:02 - 00:07:53:20


Yeah. For sure. I mean, you know, Bing started it first, as far as I know. And now Google, you do a search and the first thing that comes up, the first result is, you know, an AI result, it's giving attribution. And I forgot what we were looking at the other day. But, you know, my wife and I were searching something together, and the Google AI came back and it was just completely wrong.


00:07:53:20 - 00:08:08:28


And, you know, I feel fortunate that we are able, when we're searching some topics, to have enough familiarity to know when something's wrong, as opposed to someone else who might not. And I think that's kind of the, for me, the biggest concern about AI right now.


00:08:09:00 - 00:08:29:09


Yeah, they made a huge, I mean, I don't know why, you know, Google's stronghold has always been their reputation and their trust factor as it pertains to getting the right answers. I don't know why they pushed so quickly. Probably in response to feeling a little bit flat-footed in terms of all this AI revolution, but they pushed that out way too fast.


00:08:29:11 - 00:08:48:28


You know, I saw some results where it was like, ‘Eat three rocks a day’ and it's citing the Onion and you're like, well, the Onion is not really reputable source and sometimes is more humor-based. But, you know, I often tell people, look at Perplexity AI. It's a free AI search engine, basically. It uses search, you know, Google search results and index pages as a way to garner information,


00:08:49:01 - 00:08:54:29


and it uses ChatGPT to kind of help analyze it. But it's really powerful and amazing.


00:08:54:29 - 00:09:15:11


Yeah. So, Kathleen, speaking of Google and these other, you know, tech companies that are diving into AI, it seems like we've had some interesting case studies when it comes to good PR and probably, frankly, most bad PR when it comes to AI. So, I know you and I, before we press record, were talking about some of that.


00:09:15:17 - 00:09:21:13


Let's dive into a couple of those things that have happened, and let's just talk through those for a minute.


00:09:21:16 - 00:09:38:07


Yeah. I mean, I think Google is a prime example of that, right? You know, whether it's eating three rocks a day to, you know, ‘How do I make sure that I get the cheese on my pizza to stick?’ and it recommending, you know, Google search review, recommending that you add glue to your tomato sauce to have your pizza, you know, have the cheese stick.


00:09:38:07 - 00:10:08:17


Right. You know, I think, you know, Google has, there's such this race and need to be first that people are throwing out the idea of quality altogether. And I think it's actually, you know, on Google’s side, you know, I think they're struggling with not only a product issue, but a PR issue, like they have so much, especially in the search world, so much validation and trust to the overall public that they've completely tarnished by just trying to be to the market faster than OpenAI.


00:10:08:17 - 00:10:30:25


And, you know, Bing has done some integration with search. Obviously, Altman has been teasing its search capabilities in its own AI search engine but they've done a little bit better of a job on the rollout, if you will. Now, that's notwithstanding the whole Scarlett Johanson Sky voice controversy that has occurred.


00:10:30:27 - 00:10:53:11


Yeah, I want to get into that. And you're talking about Sam Altman, though, right? And there's a whole nother kind of PR crisis issue that happened with him, right. And so I don't know if we'll have time to get into that, but I just, I want to point out that, you know, these are just these moments in the rush, like you said, of getting it out and forgoing consistency and reliability.


00:10:53:18 - 00:11:16:14


And it's kind of reminding me of a lot of the attitudes that have happened, you know, in the last 10, 15 years in journalism, right. This idea of having to be the first to tell a story, and the accuracy is kind of thrown out the window, and you kind of look at it like, you know, the Venn diagram of, you know, do you want it good, fast, or cheap because there's no real intersection where all three get to meet.


00:11:16:16 - 00:11:25:20


Yeah, it's so true. And it's happening now. And it's funny because, you know, obviously the tech world has always kind of had a little bit of fake it til you make it kind of mentality.


00:11:25:27 - 00:11:29:04


Yeah. Or what's the thing, move fast and break stuff.


00:11:29:11 - 00:11:47:03


Yeah. And, you know, that is true. But there are times, I don't know if you remember, there was an AI pen that was released that was supposed to be able to capture and do multimodality. So you could be looking at something and ask, what is it looking at? and be able to respond. But there were so much hardware issues with it,


00:11:47:05 - 00:12:16:24


and they rolled it out so quickly to users that it completely failed, and there was a lot of negative backlash. And so, you know, I hope a lot of these AI, you know, one of the things that concerns me about this is, you know, somebody who's really passionate about what the positive potential impact can be of AI, is do we let this ego race to be first end up hindering our ability to have a public that trusts and aligns with AI and doesn’t build this resistance?


00:12:16:24 - 00:12:43:14


Right. I think, in an interview with Sam Altman, I believe it was with Lex Freeman, he was talking about AI and his concerns of AI being perceived similar to that of nuclear power, right, and how well nuclear power is for the most part pretty safe. There's such a negative perception around it, and it really halted any innovation after the Three Mile Incident because people were so afraid of it.


00:12:43:16 - 00:13:09:09


And so, you know, in that interview, he talked about how he's intentionally holding back some of the releases to make sure that the public is ready for it. And so, you know, I think, I really hope, you know, these tech companies A) start hiring PR agencies like yourself to really understand how to navigate this but ensure that we don't end up with this backlash that doesn't allow us to achieve the full potential of AI and what it can mean for society.


00:13:09:09 - 00:13:24:17


So, Kathleen, I want to dive more into that topic in just a moment when we come back on the other side from this quick break. But, for our sponsors, I also want to just say that I'm excited to have you here again because Kathleen, if you can't tell, is my trusted adviser when it comes to all things AI.


00:13:24:24 - 00:13:35:01


And I'm looking forward to diving into some of her recommendations for how you might also use AI in your public relations department and communications team.


00:13:35:01 - 00:13:59:19


You're listening to On Top of PR with your host, Jason Mudd. Jason is a trusted adviser to some of America's most admired and fastest-growing brands. He is the managing partner at Axia Public Relations, a PR agency that guides news, social, and web strategies for national companies. And now, back to the show.


00:13:59:19 - 00:14:09:22


Hello and welcome back to On Top of PR. I'm Jason Mudd with Kathleen Perley, and we are talking about AI and really having a good conversation here. As we bumped into the break and now we're back from the break,


00:14:10:00 - 00:14:36:14


Kathleen, I had mentioned that I was kind of eager to get into some of the concerns of AI and maybe even some of the, you know, proper fair uses of AI and specifically, you know, I just want to give our audience a quick tip, and that is that if you haven't started integrating AI into your crisis communication and crisis communication plan, now would be a great time to start thinking about that.


00:14:36:21 - 00:14:55:15


We recently used AI to kind of come up with some crisis scenarios then and, you know, use that as like an additional member of our team for the brainstorming effort of identifying potential crisis. Then we also used it just to kind of, you know, expand upon some of our thinking and see what else it might come up with.


00:14:55:15 - 00:15:16:26


And, you know, 20% of it was junk. The other remaining 40% was probably, you know, interesting, but we already had it covered. And there was probably 20 or 40% that was like, ‘Oh, that's an interesting consideration. Maybe we'll think about that.’ But I would look at your crisis communication plan and activities as one:


00:15:17:02 - 00:15:34:26


What could we, how could we leverage AI to benefit this? And then two: Be thinking about what happens if you have a possible data breach or something somehow, you know, gets shared on AI that you wish didn't and how do you respond to that? So, Kathleen, any words of wisdom there?


00:15:34:28 - 00:15:56:13


I mean, I would say, you know, there's a lot of ways AI can be really beneficial as it pertains to your crisis communications plan and your comms plan in general, right? I mean, I think, you know, one of the interesting things, having worked in health systems for a long time, you know, we would try to predict what would potentially come through on social media and how we would want to respond and get pre-approved responses ready to go.


00:15:56:16 - 00:16:24:16


And, you know, there was a time literally where a guy stole mercury and then ended up in the hospital with mercury poisoning, brought the mercury to the hospital, the hospital got shut down. So nothing you could have ever imagined, right, in terms of something that would possibly happen. And so the ability to utilize AI to help think of scenarios that your organization might need to be prepared for, or even I've been using it a lot for sales training or, if you want, even speaker training, right?


00:16:24:16 - 00:16:44:20


The ability to say, ‘Hey, act as a news outlet and interview this candidate on this type of topic.’ Go back and forth and take on a role of a troll, even, just to really kind of help people get ready for those types of engagement, I think, is key. But I also think, you know, you also have to be ready for the other side of it, too, right?


00:16:44:23 - 00:17:04:01


I don't remember which university did it, but they actually utilized AI as a way to send communication to their students about a challenge that was happening and something that was facing all the students. And what they did is they used AI, which was great because it helped them gain a really good perspective and a really empathetic voice for their students.


00:17:04:01 - 00:17:40:00


But the problem was they actually copied and pasted it so that you could tell it was like you're looking for this type of response colon. And then it put the response in. So you knew they were using AI. And so I think one of the things you have to be careful about is A) how you use it, making sure it still sounds like your organization, but also making sure that you are thinking through the emotional or the empathetic impact of, okay, let's say, you know, it's one thing to say, for example, ‘We used AI to develop a response for an individual who has a canceled flight.’


00:17:40:02 - 00:18:08:06


Totally fine, right? Now, ‘We used AI to draft a message to a pet owner after the deceased, after their pet has deceased at the clinic,’ for example. Probably not the best time to use AI, right? There's this idea of, like, we still want to have that human connection. And so I think the more and more we can leverage AI to assist the humans or assist, enable us to be more human-like, the better.


00:18:08:06 - 00:18:20:12


It's, you know, think about how do we take the mundane, the rote out of our lives so that we can spend more time doing what we do best, which is developing these relationships and human connections?


00:18:20:15 - 00:18:51:09


Yeah. Totally agree. That's good food for thought there as well. And just being empathetic and having the emotional intelligence of knowing when and what would be a good fit for AI versus keeping it still very personal. Just to share for our audience, you know, this week, or in the last kind of, you know, seven days, I personally have used AI to help create some interview questions that we want to ask potential employment candidates.


00:18:51:09 - 00:19:09:25


Some questions we want to ask members of leadership teams at new clients that we're working with. We used it a couple weeks ago to, you know, we inserted the key messages for one of our clients and said, ‘Hey, you know, give us 40 interview questions so we can do some mock interviews and spokesperson training from that.’


00:19:09:28 - 00:19:26:07


And I think that's just been incredibly helpful, you know, to have those tools and techniques. And so, you know, in a matter of minutes, we had questions lined up to ask the C-suite at a new client we just signed. And, you know, the saying I keep hearing is, you know, ‘AI is a good beginning and a terrible ending.’


00:19:26:07 - 00:19:42:05


So, yeah, it was a really good first draft for us. And then we went in and we made tweaks and things like that, but, you know, super helpful. And my biggest fear right now is it just produces so much content so quickly that it almost takes just as much work to go through and improve it and fix it,


00:19:42:07 - 00:20:03:00


but that first draft is awfully nice to have most of the time. And, just another thing I saw this week was a couple of articles went live for, how do I say this for our clients, you know, featuring our clients. And I asked one of our copy editors, ‘Hey, could you put this in one of the AI,


00:20:03:01 - 00:20:26:23


you know, checkers that we use to see how much of this was AI?’ And you know that these AI checkers are not perfect by any means. But they came back and was like, you know, 75% AI. And then there's a quote from our client, like, inserted into the article. And I'm wondering, is that the 25% that's not AI, was the quote that we submitted, you know, to their request, or what?


00:20:26:23 - 00:20:43:17


And then I just and I literally asked the reporter, I just, I couldn't help myself, and I didn't want to make the media relations professional on my team who was involved, I didn't want to make them ask. So I just said, ‘Hey, you know, I work here at Axia. This is our client. I saw the article you wrote.


00:20:43:19 - 00:20:59:02


I'm just curious. It seems like a lot of it was AI-generated. Are you concerned about that from a search engine ranking standpoint? From a credibility standpoint, you know, that kind of thing?’ And they were just like, ‘Oh, no, it's going to be fine.’ And I'm like, ‘Okay, it's your domain, your site, you know, whatever.’


00:20:59:05 - 00:21:08:13


And the site does have a pretty good domain authority. So it'll be interesting to see if they have any impact on that down the road. Kathleen, what would you advise them if they were coming to you for advice on that?


00:21:08:16 - 00:21:27:11


Honestly, it's one of the things I keep hearing over and over again, that if you use AI to help generate content, that you're going to start to get kind of devalued in the algorithms from a search engine ranking perspective. But I'm not seeing that be kind of come true. And I think the focus is more on quality and less on who is the creator or the origin.


00:21:27:11 - 00:21:50:13


Right. So if and what I found is, you know, when we use both humans and AI, it usually produces something even better than just a human or just AI alone. Right. And so to me, what I'm seeing is the AI tools, when they're generating content, it’s not necessarily a negative thing, it's just making sure that we're able to find value in it from a human perspective or consumer of that content.


00:21:50:13 - 00:22:01:08


As long as producing still valuable content information, I think you need to start to care less and less about how it's generated in a way.


00:22:01:10 - 00:22:26:17


Okay. Got it. Well, we have some more time left, and I would love to just kind of hear you do a round-up of, you know, if you were working in a corporate comms department or doing public relations for a national brand, what would be some of the tools that you would want to make sure you and your team are exploring and considering and maybe even moving from freemium to paid subscriptions on?


00:22:26:19 - 00:22:43:13


Yeah, yeah. I mean, honestly, the first thing I would say is this, you know, if you're sitting here listening to this and you're thinking, ‘No one uses AI in my organization, we have a strict policy,’ something that nature, the truth of the matter is they are. They might not be using it on your computers, but they are using AI technology.


00:22:43:13 - 00:23:02:00


And so I oftentimes recommend starting with a survey. Tell them they will not be penalized for anything that they share in that survey. Ask them which tools they're using. Right. You might find a great tool to use because one of your employees is already using it on their personal computer. So start with a survey to get an idea of who's using what


00:23:02:02 - 00:23:22:27


and allowing your organization kind of set some parameters about how these tools can be used moving forward. And then I think, you know, some of my favorite tools, is really focusing on what I'll call your core models or your core frontier models. So I use ChatGPT almost every day. I also use Claude, which is Anthropic’s AI tool.


00:23:22:29 - 00:23:48:21


I feel like Claude is a little bit better, from a voice-matching perspective and voice guideline-matching perspective. But it has some limitations, so it doesn't do image generation or anything like that. I typically will even ask Claude to ideate some header images for articles I'm writing, and then I'll take those outputs and put them into ChatGPT-4o to get Dall-E to kind of generate those images for me.


00:23:48:21 - 00:24:12:14


So I use those really in tandem every day. I do pay for the team version of ChatGPT and I do pay for Claude. They do not have a freemium, if you will, but they're really inexpensive. I think I pay about $25 a month. I have two team members on it. One is my dad, actually. And because I didn't want to have my data be something that it was training on.


00:24:12:14 - 00:24:34:19


So if you're on the team version, you have control over how your data is being used. It's not being used to train the general models, so you're a little bit more protected on that end. And obviously Anthropic is really heavily focused on, you know, ethics and security. So their data is, you know, your stuff that you're putting in from an input perspective is not being used to train any of the models either.


00:24:34:26 - 00:25:01:18


Actually, in the last like, I think, 48 hours, some people from the Super Alignment team from OpenAI have now moved over to Anthropic. So the Super Alignment team is focused on safety and security. So to see those moves gives me a really good indicator that Claude and Anthropic will be like the secure business ethical AI platform that you use.


00:25:01:20 - 00:25:27:00


And then the other tool that I use every day is Perplexity, which is free, Perplexity dot AI. And quite frankly, after a Google overview complete flop, Google should look at buying them. They do a fabulous job at analyzing search results, giving information back. It's great for any type of research. So if you are looking at speaking to somebody and wanting to do some industry research, it's a great place to start.


00:25:27:03 - 00:25:47:24


And then I use Midjourney a lot and Midjourney is Dall-E on steroids, so those are kind of my core four that I'm using every single day and that I have what I call like my annual subscriptions to you if they do have a paid version. And then I use some tools on a month-to-month basis just because I know it's going to shift and change.


00:25:47:24 - 00:26:14:05


And, oftentimes, you know, I use Runway a lot, which is a video AI tool and Pictory or Pictory dot AI, which is kind of like the cheaper version of Runway. But honestly, we've seen a lot around Open AI and Sora and so, you know, as that starts to get integrated into the team’s account, do I look at reevaluating what I'm using that AI technology for and kind of decrease my subscriptions there?


00:26:14:05 - 00:26:35:22


So I think, you know what I would say is spend time in it. Don't be blind to the fact that your teammates are probably using it, and figure out ways where you can leverage AI to take out the mundane of your tasks and allow your team to upskill and focus on the things that really, truly separate your organization from others.


00:26:35:25 - 00:27:05:28


Yeah, that's really good. Kathleen, thanks for sharing all that. I'll just mention to our audience that if you didn't capture all those names, we'll be sure to put the links into the episode notes, one. Two, we're also going to provide a link to a previous episode and the resource link from it, which was a sample AI policy that our agency uses that you're welcome to copy, paste, and replace our name with your name, and then, obviously, tweak it to your voice and your policies and your preferences.


00:27:06:05 - 00:27:26:17


And then, of course, have your legal adviser just look it over. But every one of our employees signs an AI ethics, AI principles, AI, acceptable use type policy at our agency. We've communicated this to our clients that we use AI and, you know, if they have any issues or concerns, that they should let us know.


00:27:26:19 - 00:27:44:16


No one has let us know that they have any issues there. Of course, we're very careful about safeguarding their information. And then, lastly, any of our vendors, we ask them to also agree to our AI principles and that they have to alert us that, you know, AI was used on this and if it was used because that's important to us as well.


00:27:44:16 - 00:28:07:21


We don't want to be surprised when we see something seems AI in it and it is or is not. So we want to give them the opportunity to disclose that upfront. And, yes, you know, you just talked about Pictory, and I'm really pleased with Pictory. I started checking that out based on your recommendation. I prepaid for a year, and then you're like, ‘Oh, you probably should have waited till whatever,’ but that's just how things go, right?


00:28:07:21 - 00:28:23:18


And yeah, I think it was already at least 20% off, you know, normally if you buy it for a year and they sent me a promo thing that made it basically be like 60% off, and I thought, let's just buy it. And, so we got it for a year and I'll tell you, our team loves it.


00:28:23:18 - 00:28:44:18


And I'm sure our clients love it. They probably don't know we're using it because we're just producing content that's video. But, you know, it's not perfect. There's some things I'd like to see a little bit different. I will say that they are very open and receptive to our input and have actively solicited it, which is refreshing because a lot of times software and other companies, they just think they've got it all figured out.


00:28:44:18 - 00:28:45:09


So yeah.


00:28:45:12 - 00:29:06:10


Well, and I don't think we're gonna see Sora time before the election, so I think you're good from that perspective. I think OpenAI is very cognizant that it could cause some challenges as it pertains to deepfakes and things of that nature. So, yeah, I don't think we're gonna see it anytime soon. And I highly doubt that they will not increase the fees for your OpenAI accounts


00:29:06:13 - 00:29:08:07


if you get that added in.


00:29:08:09 - 00:29:24:25


Yeah. Well, I want to go back to, because we didn't really hit on it, but we talked about it offline a little bit, I also want to give you an opportunity to share anything else you want to share with our audience, but we were talking earlier about the little voice issue that happened recently. And so I’d love for you to kind of update the audience on what that was


00:29:24:25 - 00:29:28:01


and then just kind of share your thoughts on it, Kathleen.


00:29:28:03 - 00:29:50:13


Yeah. So it's interesting. So, two things that has kind of come through the news, right? ChatGPT did a big release about 10 days ago on 4o, and they highlighted some of their new AI voice agents that allow you to kind of really talk and have a two-way conversation, a dialog. One of those voices, Sky sounded very, familiar,


00:29:50:16 - 00:30:16:04


and it's because it sounded very similar to Scarlett Johansson's voice, particularly her voice in ‘Her,’ which is that movie for that gentleman developed, I think it's Orlando Bloom, develops a romantic relationship with his computer. And we've, there's been a lot of news and kind of behind-the-scenes things around Sam Altman, his infatuation with that movie.


00:30:16:06 - 00:30:27:24


And when it came out, Scarlett Johansson had mentioned not only does it sound like her, but that OpenAI had reached out to her, including Sam Altman, on multiple occasions.


00:30:27:25 - 00:30:28:28


Multiple locations, that's right.


00:30:29:01 - 00:30:55:01


Yep to be the voice of ‘Her,’ or to be the voice of OpenAI and to be one of the AI agents, and she turned them down. And so that's where this really gets interesting. And so, you know, as this starts to unfold, we've seen more and more information that has come out. One particular thing that's been interesting, they actually hired a voice actor before they even reached out to Scarlett Johansson


00:30:55:03 - 00:31:30:11


who did all the recordings. That agent has come forward and said, you know, ‘This is my voice. I’ve gone through this.’ But I'm not surprised that it sounds a lot like Scarlett Johansson. I wouldn't be surprised that they didn't, you know, that they didn't intentionally look for somebody who sounded like Scarlett Johansson. Now, the agent has been, or the agent of the voice actor has shared that she was not given any direction to sound like Scarlett Johansson or mimic the voice in ‘Her,’ but it really calls into question, and it's something I think is really interesting with


00:31:30:12 - 00:32:01:22


my background in linguistics is how similar voices can sound and how our auditory capacity to really translate just the most minute variations in sound and whether it's inflection, tone, or anything of that nature, how it's hard for us to tell the difference. And so it isn't Scarlett Johansson's voice. It wasn't trained on her voice. The voice actress that was used for that voice was not given direction to sound like her.


00:32:01:24 - 00:32:32:29


But it really just calls into this thing where there's no legal issue here, really, based on what we've seen. But it is really an ethical question, right? And so, you know, it's one of those interesting things to see how it will unfold. Obviously, OpenAI AI has come out and said ‘We've taken down the voice out of respect for Scarlett Johansson,’ but how do we kind of thread that needle, if you will, in terms of having voices that seem approachable, that can seem very human-like and imitating deepfakes,


00:32:32:29 - 00:32:52:03


if you will? And so I don't think there's any real legal challenges here. Obviously, the fact that Sam Altman, I think 24 hours or on the day of or very closely to the release of the Sky voice, made a comment about, you know, tweeted the words ‘her’ at one point


00:32:52:05 - 00:32:52:19


That's right.


00:32:52:25 - 00:32:59:00


isn’t helpful in terms of their case, and this is where they could use a great PR agency to help them navigate that.


00:32:59:00 - 00:33:00:10




00:33:00:13 - 00:33:23:03


But, you know, it's one of those things where I think they really need to stop, you know, playing so fast and so loose on some of these things and really kind of pay attention to it. But I also think as a consumer population, we need to, you know, innovation sometimes is going to be messy. And I just really hope it doesn't derail us, if you will, in terms of achieving what's really possible with AI.


00:33:23:06 - 00:33:46:04


You know, one of the things that my students at Rice were working on, who I was really amazed with, we used, a tool called Soul Machines, which is an AI avatar. I was working with her on integrating it with Hume, which is an API that makes it sound more human-like and empathetic. But we were working on building a AI sponsor for individuals who are struggling with addiction,


00:33:46:06 - 00:34:09:22


and there's so much huge potential for that. And you want that voice to sound real and human-like, right? And so, you know, I think this Scarlett Johansson thing is a challenge that we're going to have to face from an AI perspective and navigate from an ethical consideration standpoint. But I just am hoping that it doesn't completely derail us from what is possible in the positive impact we could have.


00:34:09:22 - 00:34:12:07


and see with AI.


00:34:12:10 - 00:34:36:25


Yeah, it's going to be interesting. An ethical case study, a business case study, a technology case study, a culture and impact and ethics case study, it’s just, it's all these things, which make it very interesting. But, you know, at the end of the day, you know, what we don't want is technology to replace the arts. We don't want technology to replace, you know, the ability for people to earn a living and all of that stuff.


00:34:36:27 - 00:34:57:04


At the same time, we are looking to build a better, faster mousetrap. And at the end of the day, it's going to be interesting to see what happens and the shift in our employment and economics and all other things and even political systems, perhaps. So yeah, you know, this is going to be big. And, you know, I hope that there's never a company like Skynet we got to worry about and all that fun stuff.


00:34:57:08 - 00:35:16:27


Yeah. But, you know, or even like the movie Surrogates, which, you know, is, you watch that under the lens of AI in the pandemic, and you start to wonder, like, what did they know that we didn't know, you know? But, yeah. Kathleen, real quick, I would just ask you for if somebody wants to dive more into this topic,


00:35:16:27 - 00:35:39:00


one, how did they get ahold of you? Two, what are some things that you're reading and consuming, you know, and following that they should be checking out, whether that's books or, you know, YouTube channels, or social media influencers? And then, of course, I would love to give you an opportunity, if you can do all this at once, to plug your newsletter.


00:35:39:03 - 00:36:03:16


Yeah. So, for things to see if you want to dive deeper into it, I would say start testing and playing around with the tools. That is probably the easiest and best way to really start to feel more comfortable, gain more knowledge. There's a course on Coursera called AI for Everyone. It's by Andrew Ng. I would also recommend, in terms of who I follow and read,


00:36:03:19 - 00:36:27:16


Paul Roetzer with the Marketing AI Institute has a wonderful podcast that comes out every Tuesday that is really incredible. I also love to follow a lot of the leaders in the space. So Sam Altman, Ilya, a bunch of the kind of CEOs of a lot of these AI frontier models, because their interviews are often the most telling in terms of where the industry is going.


00:36:27:19 - 00:36:46:25


And as Jason mentioned, I actually have a newsletter that I send out every Thursday. I built it initially for my MBA students who we started the class with a rapid fire every week. And now as they've graduated, they've asked that there's a way to stay on top of it, so I now kind of do this homework for myself and share with everyone else.


00:36:46:25 - 00:36:55:11


So if you'd like to subscribe to that newsletter, I'll give Jason the link and he'll put it in the notes for this podcast.


00:36:55:13 - 00:37:19:21


That sounds great, Kathleen. As always, and I was not surprised you were a great guest today. I appreciate you sharing all of your smarts here, and I'm happy that we have a relationship where I can continue to live vicariously through you and learn from you and hear all the exciting things that are going on and, like I said, you just being my trusted adviser on how do we best do all these great things that are happening.


00:37:19:21 - 00:37:45:10


And I guess I'll just end with this question. Kathleen, you've been in the agency space. I'm in the agency space currently. You know, if somebody's listening to this and going, ‘Gosh, I wonder what my agency does for AI?’ or maybe they even know ‘My agency's not even touching AI.’ Kind of what would you recommend that person sitting in that seat, being a buyer of or decision-maker of PR agency services?


00:37:45:10 - 00:37:53:25


Is that something you think they should be concerned about, or should they even be seeking more transparency from their agency in their current AI position?


00:37:53:27 - 00:38:17:21


I think they should 100% be seeking transparency, right? Agencies, whether it's marketing, advertising, or PR agencies, to have a true partnership, you have to have transparency. We always talk about, you know, at my old agency, agency, we used to say, we treat you like family, which means we'll tell you if that dress looks bad on you. Right. And so you need a partner who's willing to share the good, bad, and the ugly in terms of leveraging new technologies.


00:38:17:24 - 00:38:43:25


But these technologies can really assist with making sure that you have the subject matter experts focusing their time, energy, and talent and your dollars on the things that matter most, the things that will add the most value. And so if they're using AI to take some of the behind-the-scenes grunt work off their plates, I think it should be welcomed and embraced because you guys will be getting, you know, organizations will be getting a better deal out of it, from that perspective.


00:38:43:25 - 00:38:51:06


So, but that's all reliant on making sure you have transparency and communication between your partners.


00:38:51:08 - 00:39:07:10


Yeah. We treat AI like, you know, a team member, right? So, you know, we stole this from an agency that you and I both know, but we call it Janet from The Good Place. Right? So we run things by Janet. Make sure Janet is, you know, involved in something. And we protect clients' confidential data.


00:39:07:10 - 00:39:27:04


We use, you know, and, you know, we describe the industry they’re in and then we make up a company name and then, you know, just to make sure that we're safeguarding as much as possible. Nothing confidential is put in it. But I think that's a good way to think about it. And, you know, pardon the cliche, but make sure AI has a seat at the table for brainstorming meetings,


00:39:27:04 - 00:39:48:29


for planning meetings and things like that. And you can actually assign somebody and maybe it's, you know, a low-ranking person or maybe somebody who just isn't very participatory in these types of meetings, and they could just be the AI, you know, you know, engineer, the person giving the input to the AI, and then we can say, ‘Hey, well, what does the AI think?’ or something like that.


00:39:48:29 - 00:40:02:19


So, you know, just another way to kind of think about it. And, you know, sometimes the responses that it gives you are, you know, at or below what you'd expect from an intern. But then sometimes there's a nugget in there that's just extremely valuable.


00:40:02:21 - 00:40:28:29


Yep, 100%. We actually, there was a study at Rice that was done that I thought was really fascinating. They actually used the quarterly, like financial briefings. And once it got to the question part of the quarterly financial briefings or overviews, they actually asked the same question that was being asked to the CEO. They asked it to AI, and they identified the closer were, the closer the answers were between AI and the CEO or the CFO,


00:40:28:29 - 00:40:50:11


whoever's reporting the quarterly financials, the more likely the stock would go up and the performance of the company would go up. The further apart those two answers were, the stock would go down and they would short it, from that standpoint, or they would predict to short that stock. And their accuracy was like above 90% in terms of the predictions.


00:40:50:11 - 00:41:10:07


And so it's really fascinating. And I think, you know, oftentimes people will say, you know, ‘I can't get that good of results out of AI; I might as well do it. I'm faster at it.’ A lot of that is user error. And the more that you can kind of continue to test and use and give feedback to the, you know, reinforcement learning, the better these tools get to enabling you.


00:41:10:10 - 00:41:31:16


I totally agree with that. And, the one example I give when I'm asked to speak about AI to industry associations or companies is I tell a story. When I was an intern, I worked at a PR agency, and they were hiring a third party to build a PowerPoint deck for one of our clients. And I'm there's as the college intern that doesn't have any credibility at week one.


00:41:31:16 - 00:41:49:16


And I'm just like, I could do that. Why are you outsourcing this? And at the time, I think they're paying like $125 a slide. And in 1996 or 7, that was a lot of money. And I'm like, ‘Guys, I could do this.’ They’re like, ‘Oh, no, we picked this firm. They specialize in PowerPoint slides.’ ‘Okay,


00:41:49:16 - 00:42:08:20


I've been doing PowerPoint slides for years, you know, and they're like ‘Oh, well not at the level of what we're looking for.’ Well then when that PowerPoint agency or whatever they were, audio-visual company basically ghosted them and was not delivering in time, suddenly they're like, ‘Hey, Jason, you still want to do the slides?’ you know, whatever.


00:42:08:26 - 00:42:23:28


And I knocked them out and like a day. And they were just blown away. And good news is my stock, you know, increased in the organization because of that opportunity. And the people on the digital side wanted me on their side. But at the end of the day, I look at that as like AI today like it.


00:42:24:01 - 00:42:45:17


People think that it's so early and all this stuff, but really the input you give it is often what you get back, and sometimes you just have to rephrase and say, ‘Hey, let me rephrase this input I gave to you and instead look at it this way.’ Because we talk as humans a certain way and, you know, sometimes we say something and it's not really what we mean, but we're able to perceive or understand what the other person is trying to say.


00:42:45:20 - 00:43:04:27


Yep. So what's interesting, when we start to talk about multimodal and really utilizing body language in addition to some of the things like tone and kind of the velocity of which speech is happening to identify like kind of emotions, but be able to start bringing in body language in terms of AI, understanding what you're trying to get across.


00:43:04:29 - 00:43:23:12


Yeah, for sure. Well, Kathleen, thank you so much for sharing today. I'm really glad that we did this. It was great seeing you recently in Denver for the AMA summit. And with that, we're going to sign off here and say thank you to our audience for tuning in to On Top of PR.


00:43:23:13 - 00:43:39:19


That is our goal here is to help you stay on top of PR. If you found this episode helpful, I do hope you'll share it with a friend or colleague who would benefit from it. And if you have any requests on topics we might explore in the future, please feel free to drop me a note or comment on our social media posts.


00:43:39:24 - 00:43:45:19


We love requests and are always happy to honor them. And we want to thank our guest, Kathleen, for being here. Please check out those episode notes for links and further information. And with that, thank you so much for tuning in and be well.


00:43:45:19 - 00:44:41:10


This has been On Top of PR with Jason Mudd, presented by ReviewMaxer. Be sure to subscribe so you don't miss an episode, and check out past shows at ontopofpr.com.

Axia PR logo. ReviewMaxer logo.



Jason Mudd's image

About your host Jason Mudd

On Top of PR host, Jason Mudd, is a trusted adviser and dynamic strategist for some of America’s most admired brands and fastest-growing companies. Since 1994, he’s worked with American Airlines, Budweiser, Dave & Buster’s, H&R Block, Hilton, HP, Miller Lite, New York Life, Pizza Hut, Southern Comfort, and Verizon. He founded Axia Public Relations in July 2002. Forbes named Axia as one of America’s Best PR Agencies.


Find more On Top of PR episodes on: 








Apple Podcasts


iHeart Radio




Google Podcasts

Amazon Music






Topics: On Top of PR, artificial intelligence

Liked this blog post? Share it with others!


Comment on This Article

Blog Subscription

Recent Posts

Popular Posts