Episode 335 - AI

Transcript:

Pete: Hey, Jen.

Jen: Hey, Pete.

Pete: So I feel like we've talked around this topic in many different ways over the last twelve months.

Jen: Mmm?

Pete: And I've probably reluctantly done so, because I'm super mindful of not getting onboard buzzy buzz-words.

Jen: Okay.

Pete: But I think it's worth us discussing and talking about. And just for my own curiosity, I want to know from your perspective, how are you using AI tools in your day-to-day?

Jen: Ooh.

Pete: That's my question for you.

Jen: That's juicy. This is The Long and The Short Of It.

Pete: So this might be a personal quirk of mine, but I tend to push back on buzzy buzz-words.

Jen: Mmm-hmm.

Pete: And not want to oversaturate the podcasting landscape in this case, because, gosh, how many podcasts and blogs and articles and books are now being written about AI, AI, AI? It feels like it's everywhere.

Jen: Because it is.

Pete: And the more conversations I am having with people about the ways they are and the ways I am using it, the more I'm realizing, in so many contexts, we're all using it without really talking about it.

Jen: Yeah.

Pete: Even though it feels oversaturated. And so, I just think it's helpful to hear from other people. And I want to hear (selfishly) from you, as to how you are thinking about and/or using already AI tools. This is like just a, literally a practical, "How do we use these tools effectively in our worlds," kind of episode. What do you think?

Jen: Well, I guess the first thing to say is, I was using AI without knowing it was AI for a while.

Pete: Right. Say more.

Jen: Well, even things like my scheduling app and auto-respond and things that are just sort of like...they're just in the background of our everyday lives. But now that we're talking about generative AI and language models, it feels like a little more, I don't know, "man behind the curtain". Like, we're pulling back the curtain and being like, "We're actually using AI now," as opposed to like, I didn't realize it was sort of baked into my life.

Pete: Yeah. We're intentionally choosing to use it.

Jen: Yes, to create things, to generate things.

Pete: Which, not everybody...so this already, I feel like you're getting into your little mindset or philosophy for thinking about using these tools. And you're right, I guess we are thinking about or I'm thinking about large language models and generative AI. How do you think about using these tools? Do you like these tools? Are you against them? Are you for them? Do you see a world where your clients are using them really effectively? I just want to know.

Jen: Oh, yeah, I use them a lot. Lately, I've been bouncing back and forth between ChatGPT and Gemini. And I don't know that I have a preference. I just sort of am interested in what one will tell me over the other.

Pete: Right.

Jen: In very broad terms, I use AI for research. I use AI for extracting thematic content. I use AI for summarizing my brainstorms. I use AI to give me creative prompts when I'm stuck. I use AI, I will rattle off like big ideas and then turn those ideas into an agenda for a meeting. Sometimes I will use AI, I will import a transcript and ask the robots to tell me things about myself that I might not have previously known.

Pete: So good.

Jen: I use it a lot.

Pete: It sounds like a companion of yours or a co-pilot, in a way, is like how you're thinking about it. It's not limited to, "This is the one specific use case that I'm using it for." Like for example, in this conversation, we have enabled the AI recording summary in Zoom, and so that would be one really narrow use case that one could use it for, is like transcribing meetings. But you're using it for so much more than that.

Jen: I use it for so much. And by the way, the Zoom AI companion is hilarious for some of my coachings. Because, you know, we do a lot of scripted material. And so, if you turn it on, it like is so confused. But if I'm doing a strategy session with someone, I tell them, "Don't worry about taking notes. This is going to summarize the conversation and send you a list of to do's." And it does. It does it in like less than sixty seconds.

Pete: Totally. Do you find that enables you and the client to be more present or more focused or more connected to the actual task?

Jen: Yes, because we're not afraid of losing the thread or losing the thought. And you know, sometimes when you're in a strategy moment, you're wanting to capture all of the action items. And what I love about the Zoom AI companion is it turns things into action items that I didn't even realize I was assigning as an action item.

Pete: Right, right, right. Okay, this is juicy. Like, I take from your thread...one of these conversations I had recently with a very senior executive who's fascinated and nerding out on generative AI, and she sent me this big, long email. Which, we're having this really cool email exchange. In the last paragraph, I will paraphrase it, but she said something that I just think is so awesome. She says, "A lot of the above," i.e the assertions that she made about where these technologies are going to be helpful, "will enable leaders to be more present with their team and have more human interactions with their team, which is something they all crave and lack at the moment."

Jen: Right.

Pete: And so that idea which speaks to your idea, I almost ladder up, to go, how do we leverage these tools? This is how I'm thinking about them. How do I leverage these tools to enable me to be more human, rather than to replace my humanity?

Jen: Right.

Pete: Which, I feel like is so much of the dystopian conversation. I'm thinking about, and that's why I used the word "co-pilot", how do I have a co-pilot here that helps me be more human? That takes away things like note-taking or keeping track of certain threads or processing a whole random list of ideas into a distilled list of ideas? Those kind of boring, mundane, I wouldn't say necessarily human tasks, how do I get AI to enable those tasks so that I can focus on being present and asking the right questions and, you know, being more (again) human? What do you think about that?

Jen: A thousand percent, yes. Maybe, it's helpful for me to say this. One of the things I always remind myself and my clients, because I encourage them to experiment with AI, is, this is a robot. It doesn't care about you. It doesn't care about me. It doesn't care about anything, because it's a robot. It has no feelings. And because it has no feelings, you can have feelings.

Pete: Mmm.

Jen: Like, it's not going to take on your imposter syndrome when it's doing its research. And when it puts its research in front of you and you feel the imposter syndrome, it's sometimes easier to recognize it in yourself, because this other robot that doesn't love you is like, "Here's the answer you're looking for."

Pete: Right.

Jen: And it can sort of help you confront yourself in that way.

Pete: Yeah.

Jen: So I just always remind myself, this is a robot.

Pete: I love that reminder because there's a very specific application that I've been using myself and telling leaders to use, which is to process feedback and/or give feedback.

Jen: Mmm-hmm.

Pete: So that might look like uploading a transcript of a keynote or a workshop that I was going to run and saying, "I want you to take on the role of an audience member based on the fact that this is a-," I'm making this up, "-a finance team of senior leaders at a large bank here in Australia. And I want you to think about how you would challenge or question this talk or this topic or this workshop." And it will do so without emotion because, to your point, it's not worried about, "I don't want to hurt Pete's feelings. I don't know how I'm going to deliver this feedback." It's pretty black and white, which is positive in the way that it gives feedback. But then, I think the other thing it's positive in is, it's harder to take personally. I mean, you can't take personally the feedback, because it's just like a robot generating text.

Jen: Right.

Pete: Whereas if I asked a human to give me feedback, and I did the dry run of a talk in front of them and they gave me the feedback, even though I'm pretty well-versed at receiving feedback and I like to think I have a growth mind and all those things, I could easily go, "Oh, yeah, but that's just because Jen doesn't really like structuring keynotes in that way. That's the reason she's giving me that feedback," and I immediately make it personal,

Jen: Right.

Pete: So it's like, these tools can help us remove humanity in the positive and enable us to be more human in the positive.

Jen: And then, the other thing to remember, to the example you were just giving, Pete, is that when you are getting feedback from another person, it is personal. That's the nature of the business, and why we have to cultivate more empathy, better listening, stronger communication. Because when you're giving feedback to a human, you are not giving it to a robot. You're giving it to a person who does care, who does have hopes, and dreams, and fears. Robots don't have hopes, and dreams, and fears.

Pete: Mmm.

Jen: Unless you say to it, "Have a hope, a dream, and a fear as you answer this prompt."

Pete: "And here is the hope, and the dream, and the fear." Which, is also another way that you could use these tools.

Jen: Yeah. It's a great way to prompt it, honestly.

Pete: Right. Yeah. I think so much of the tactics that I've been helping leaders with comes down to assigning roles to Gen AI platforms, whether it's Gemini or ChatGPT or Perplexity or Claude or whatever new one that you're using by the time you listen to this. It's like, "I want you to take on the role of a senior HR professional. And you are time poor. You really want to deliver a great leadership program. And you don't have the capability to spend, you know, this much money. You only have this particular budget. How would you respond to this?" So like, assigning it a role enables you to get more effective feedback or responses from the inputs that you give it.

Jen: Absolutely. You know, while we're talking about the benefits, I feel like we also have to call out some of the red flags. And Pete, I told you this story last week. It's just so funny. I was sitting on a panel, like doing a fireside chat thing. And the person who was the interviewer, a very senior leader in a large organization, had been given the script of questions to ask me. And I got to see the script ahead of time, and it was like, "You once said...," and then, it like listed all of these things that it attributed to me, and, you know, "Unpack this for us." And I was like, "Oh my god. Oh my god." I said to my husband, I was like, "I think they used ChatGPT to create the script. And now, I don't know what to do, because I never said any of these things. And like, what do I do? What do I do?" And then, finally, I was like, "Do what I would do as a human being. Reach out and say, 'I never said any of these things. Can we please revise the script?'" Which, we did. But like, ChatGPT got it wrong. It just made shit up.

Pete: Right.

Jen: So, you do have to fact check.

Pete: Right. They call it "hallucinations". It's so common, yeah, yeah, yeah. This happened to me last week. I did a whole like, "I want you to take on the role of a creative design agency, and analyze this website." And I gave them my website. "And provide feedback on what you would do to improve the positioning of this particular organization." And I gave it some more context. Anyway, it came back with this like, this list of things, like, "When you say you help people see around corners, you should reframe that as blah, blah, blah." And I was like, "My website doesn't say it helps people see around corners at all. That's totally made up." So I love that call out, because it's down to this...god, I have this separate rant that we should do a separate episode on...this skill of discernment.

Jen: Yeah.

Pete: And how that's only going to get more critically important for us, is to play around with these tools, to ask it questions, to try different experiments and roles, and to kind of hold it all pretty lightly, to take what works or what is helpful and challenge and question and disregard the rest, especially when it's not accurate.

Jen: Yes, yes, yes. One of the things that I use AI for is reducing the amount of unnecessary hours that I can spend falling down research rabbit holes, to take me quicker to the learnings that I could apply from that research. So as an example, I will have a client, let's say, auditioning for a production directed by a specific director. In the past, I would have done hours of research about this director and their full body of work. I would, you know, identify thematic content that is true across different projects. Now, I just say to the robot, "Tell me about Director XYZ. What sort of thematic content do they explore in their full body of work? And as I prepare a client to audition for their upcoming production of XYZ, what should I keep in mind?"

Pete: Mmm.

Jen: And it'll say, you know, like, "They explore the themes of family dysfunction." And if I entered a different director's name, it would say, you know, "They explore resilience," or whatever it might be. "And you should keep in mind these things that are true across their whole body of work." And you may or may not even agree with the AI's conclusion, but it gives you something to respond to. And I do believe that creativity is a response, so I'm responding to this information in a creative way.

Pete: For sure. I think it helps so much with when you have that issue with like a blank page, when you go, "Oh, I need to write a draft for, I don't know, this proposal that I'm putting together, in my context. And here are the things I know about it, but goddamn it, I just can't figure out how to start this." I'll just get a really, I mean, it's almost not even a shitty first draft, it's like a zero draft from prompting a tool like ChatGPT. And then, I'll go, "Oh, okay. This is crap for these reasons. I know how to fix this up. I know how to do it." And I'll start to iterate and improve it. And then, I might send it back and say, "How do you make this better?" And like, it becomes this dialogue again...this is why I think they're co-pilots...rather than a one and done, where it's like you ask it one question and that's it.

Jen: Yeah. And I think we've said this on this show in other episodes, but the robot's ability to be helpful is impacted directly by the quality of the prompt you've given it. So you can't say to ChatGPT, for example, "I'm stuck on a project, help me," and then, expect it to help you. You have to be much more specific.

Pete: Okay, this is a nice tee up for a rant I have about this...because I agree.

Jen: Okay.

Pete: This feels like another example to me of how these tools can help us be more human and better leaders, in the context that I spend a lot of my time helping people with. Because that example is the same with humans.

Jen: Yep.

Pete: That if I'm trying to get better at coaching people as a leader, and what I'm trying to do is ask better questions to enable them to figure out the direction or the path that we want to go, that's the exact same skill, in my mind, as prompting a generative AI to get a desired output or result. What's the appropriate amount of context? What's the appropriate prompting question? You know, we call it "prompt engineering", which is essentially being curious and asking questions.

Jen: Right.

Pete: It's the same skill that comes from coaching. It's the same skill that comes from being a leader. And it's such a great, I think, training ground or sand-pit for us to get better at asking questions. If you're too scared to start asking your team a bunch of questions because you feel like you're not, you know, as well-versed in being a coach and asking questions as you'd like to be, you can practice with ChatGPT. You can practice with Gemini. And you'll notice that the better questions you ask, the more clarity that you provide, the better output that you get, just like humans.

Jen: Oh, it's so true.

Pete: It is. And I also feel like...I'm also like mindful that this episode could be so redundant in six months time, because the use cases will be so exponential.

Jen: I know.

Pete: Which is so fun. But I have a couple of other, and I'm sure you must have some, like fun ways that you've used these tools that are perhaps a little outside the box.

Jen: Yeah.

Pete: So I have a separate project, that I've assigned the role of strength and fitness coach to ChatGPT. I have uploaded multiple different strength programs that I have accumulated over the years, that have been created with physios and strength coaches, that I've used over the years. And I've asked it to distill all of these training logs into like a step-by-step, three-session-a-week program. And I stay accountable to this ChatGPT, so I report at the end of each gym session, "Here's what I did. Here were the weights I used." And it's like, "Great job, Pete. Next time, try focusing on your technique or lifting this weight a little bit, like go a little bit heavier." And so, like I'm thinking about it in ways of trying to make myself more accountable to the things I want to be accountable to.

Jen: Ooh, I love that. That's really cool. In terms of like the accumulated knowledge, I do have some very old threads with my ChatGPT that I just keep building on.

Pete: Yes!

Jen: Where like in the early days, like for example, when I started The Career Collective (which is my online career strategy community), I could not get ChatGPT to understand what I was throwing at it. I'd be like, "Okay, I'm running a ninety-minute workshop. This is what it's about. This is how I want to structure it." And then, it would send me something, and I'd be like, "You're not listening. Listen to what I'm saying."

Pete: Right.

Jen: Now, you know, I keep that same thread, and I'm like, "Okay. New month, new topic. These are the points I want to make. I want to continue with the structure that's been working. I want to add this. I want to take away that. Give me a timed agenda."

Pete: Mmm.

Jen: And then, in like the first or second try, it gives me the timed agenda. And then, I can go in and I can judge it. But it's not based on its accumulated knowledge, it's based now on my accumulated knowledge.

Pete: Right. I love that. A friend of ours sent a prompt to me which I sent on to you, which builds on that, which was to ask the model that you've been using the most, "Based on everything that you know about me, what do you feel like is something I need to hear that perhaps I haven't heard? Can you tell me what my blind spot might be, based on all our interactions?" And I've got to tell you, the response I got back was pretty damn good. It was like straight between the eyes, "Here's what you're good at and here is what that might mean about you, and some areas for focus." It's like, woof.

Jen: I had one the other day Pete, where I was composing a coaching philosophy statement that I wanted to share possibly on my website, possibly privately. I haven't decided where I'm going to put it yet. So I wrote the statement and I put it into ChatGPT, and I said, "Make me a list of the phrases that I've just used, that any average coach would use."

Pete: Hmm, nice.

Jen: And it was like, "Here are the things any average coach would say." And I was like, "Okay, that's where I'm going to put my own special take, because I don't want to sound like any other coach. I want to sound like me."

Pete: Nice. That's great. God, there's so many use cases. I uploaded a photo the other day and said, "I want you to take on the role of...," this is so cringe, "...of world-famous fashion advisor, and provide some analysis of the outfit that I'm wearing." And it did. And it was really good. It was like, "Mmm, it's maybe...that white t-shirt might be a little too casual for this workshop you're about to run. Try a collared shirt instead." So I was like, "Alright, I'll try a collared shirt."

Jen: Oh my gosh, that is so fun. Yeah, it can be a lot of fun to play with it.

Pete: I think that's a great word. I feel like the way to approach this is with fun, is with levity, is with the acknowledgement that these technologies are the worst they're ever going to be right now.

Jen: Yep.

Pete: And they're only going to get better and more, I think, a part of our lives. And so, start playing around. Start having fun. Send us your favorite prompts and your favorite use cases: hello@thelongandtheshortpodcast.com. We would love to hear.

Jen: Oh my gosh, yes. Please send us your prompts. I cannot wait to hear how you are using this technology. I'm always amazed when I ask people if they're using AI, and they tell me a use, and I'm like, "That never occurred to me." So yes, listeners, send us your AI prompts. We can't wait to hear them.

Pete: And I'm going to include the AI summary of this podcast in our Box O' Goodies, just to make this really meta. I'm so curious what it will say. You can find that by subscribing to our Box O' Goodies, if you don't already, also at thelongandtheshortpodcast.com.

Jen: And that is The Long and The Short Of It.