Publish & Prosper
Where publishing, ecommerce, and marketing collide. Brought to you by Lulu.com.
Publish & Prosper
The Robot in the Room: AI for Authors & Content Creators
In this episode Lauren and Matt address the robot in the room – artificial intelligence! We share a few of our favorite ways to use AI and generative AI tools to help streamline content creation. Learn how we use AI as a research and analytics aid, to help with drafting and outlining content, to come up with variations on email subject lines and titles, and more!
Dive Deeper
💡Curious about any of the AI or resources we referenced in the episode? Check them out:
- ChatGPT
- DALL-E
- Midjourney
- Bard
- Semrush Writing Assistant
- Grammarly
- Brian Fanzo
- Marketing AI Institute
- Conversations With AI
💡Read Think Outside The Bot: How AI Can Help Content Creators and Using Custom AIs to Grow Your Content Business on the Lulu Blog
💡Watch Ann Handley’s CEX 2023 Keynote on How to Write Like a Robot Can't: Finding Your Writing Voice in an AI World
Sound Bites From This Episode
🎙️ [11:47] “I think the main way, right now, that most people are using it with any sort of success rate and efficiency rate is using it to buy back time in your day.”
🎙️ [15:38] ““Asking ChatGPT ‘can you break this down and explain it in a way that a room full of high school students, or take it deeper and say a room full of fifth graders, would understand,’ you'd be amazed at what you get back.””
🎙️ [41:39] “For whatever you're doing with it right now, whether that's not using it and being afraid of it, or whether that's starting to experiment with it and getting over your fears, or whether you consider yourself an expert in it right now…this is still the dumbest version of it you'll ever use, because it's changing every single day. So if you're finding value in it today, or you're gonna take some of these things we talked about back to your desk and try to make use of them, imagine if it's even the least bit helpful today, imagine how helpful it can be in your life six months from now, a year from now.”
💀 Can’t wait for our next episode? Check out our Resources page for links to our blog,
our YouTube channel, and more.
💀 Find us on Facebook, X, Instagram, and LinkedIn at luludotcom!
💀 Email us at podcast@lulu.com
💀 Sign up for our mailing list.
Matt: Hello, everybody. Thanks for joining us today for another episode of Publish & Prosper. I'm sitting here with Lauren, and today we are going to talk about the robotic elephant in the room. That's right, AI.
Lauren: I'm very excited to talk about this. And a little nervous too.
Matt: Are you?
Lauren: Yeah.
Matt: Okay. I think it'll be fun.
Lauren: I think it will be too. And we all know it's a hot topic right now. So...
Matt: Yeah. We're going to keep it fairly Primer 101 style, we're gonna talk about some of the basics. We're gonna focus on the here and now and what you can actually do with AI that will increase your efficiencies and not focus so much on what the future of AI might look like, at least not in this episode. So for now, we're gonna acknowledge the fact that AI has some potential pitfalls and/or uses that could be construed as not so productive or nice, but we are gonna focus on the positives and what it can do for you today.
Lauren: I think that sounds pretty good.
Matt: Good.
Lauren: Yeah, let's demystify the robot in the room a little bit.
[1:36]
Matt: I agree, I agree. And we'd be remiss without first talking about the fact that certainly there'll be people listening, and just people out there in general, who are not fans of AI right now. But it's important to note that most of us, in one fashion or another, have already been using AI-powered tools and/or devices, pieces of hardware, things like that.
Lauren: Really, like what?
Matt: Well, for example, I have an Alexa in my living room.
Lauren: Oh, do you?
Matt: I do. Not a big fan of it, but I have one there, and that's basically AI-powered.
Lauren: I'm Team Google personally.
Matt: There's a number of devices and/or tools and/or pieces of software that many of us use, either on a regular basis or have used intermittently, that they're powered by AI and we just don't think about it, or we don't realize it, or we don't know it, but in many ways we've all been working with AI, and machine learning, and tools that incorporate those things for years and years, and we've just never really thought about it too much.
Lauren: Sure, even things like Netflix recommending to me what it thinks I'm gonna like based on the last couple of movies or TV shows that I binged or.
Matt: But that's obvious though.
Lauren: Is it?
Matt: Yeah, knowing what Lauren's gonna like is pretty obvious.
Lauren: I thought you meant that it was obvious that Netflix was using AI.
Matt: Netflix has a hard time in my household figuring out what it should suggest to us –
Lauren: Oh I’m sure.
Matt: – But for you, I imagine it's pretty easy.
Lauren: Yeah, yeah. I can't, I actually can't say a single thing against that, that is completely accurate.
Matt: But you're a hundred percent right. And whether it's Netflix, or Spotify suggesting a new band to you, or things like that, these are all things that are indicative of machine learning and AI. And we all love these things. Nobody's running for the hills because Spotify suggested a new band to you.
Lauren: I make excessive use of the Spotify radio feature, or their new, like, smart shuffle feature where it'll recommend songs based on a playlist, mixing them right into the playlist. Those are absolutely very useful everyday examples of AI.
Matt: Yeah. And then for a lot of us – I don't want to say in the corporate world, but a lot of people in the corporate world, or even a more immediate audience, and people that we talk to every day and our content entrepreneur audiences, and creator audiences – we use tools every day that have AI, and machine learning, and some of these other things built in. Here at Lulu, we use tools like ClickUp, Semrush, and some other things that help us with our search engine optimization and our project management. And so do a lot of other companies and brands and creators use things like Jasper, and Lately, and SparkToro, the list goes on and on and on. And I'd say even just this year alone, I bet there's been well over 300 new tools that have come on the market. There's no shortage, there's a lot of tools out there that you can find and they run the gamut, everything from virtual assistant tools to things that'll help you schedule and post your social media stuff, data aggregation and analysis tools, you name it. They're all over the board. A lot of great tools out there that can help you that are based on AI machine learning.
Lauren: I don't know how I would have done my job without some of these tools that we talk about. I am addicted to my ClickUp task list. That is an essential for me. And as a social media manager, Sprout Social was essential for my job. I don't know how I would have been able to get through a single day without using a tool like Sprout Social.
Matt: Yeah, I bet a lot of social media managers would agree with you on that one.
Lauren: Yeah. So of course we've all been using tools like AI, or even tools that might be more familiar, might sound more familiar at least, for the writers versus the content entrepreneurs in the room. Tools like Grammarly and ProWritingAid are AI. Grammarly coming in there and saying, ‘hey, you used the wrong word here. We suggest this, that, and the other thing instead.’ That's AI. So then I guess the real next step that we've taken now, which is kind of where we are now, is like we were just saying, generative AI, like ChatGPT, Bard, DALL-E, Midjourney, stuff like that.
[5:10]
Matt: You just mentioned something that we should also touch on. You used the term generative AI, and that is specifically how you would refer to tools like ChatGPT, or Bard, DALL-E, Midjourney, some of those things. So at the top of the episode, we basically refer to AI, in talking about AI, but it's important to know that AI is more of a catchall term. And within the concept of artificial intelligence, there are several different things in there, like generative AI, that are slightly distinctive, or large language models, or things like that. I just wanted to clear up for anybody listening who might not be very up on the terminology around AI and what generative AI was when you brought that up
Lauren: That's a great point and a great clarification. Maybe we should take a step back a little bit and do this. I do use ChatGPT pretty regularly for work, and we'll get into some of those uses about how I use it and how other content creators can use it later in the episode. But one of the ways that I thought it would be really fun to use it for this particular episode was to ask it to define itself. So I asked ChatGPT to define AI. It gave me a very thorough definition, actually, which was nice, but some of the quotes that I pulled from it were ‘artificial intelligence, parentheses AI, refers to the simulation of human intelligence in machines that are programmed to think and learn like humans.’ But then it went on to explain to me that there are different types of AI and it categorized them as ‘narrow or weak AI,’ which is AI focused on specific tasks. And then ‘general or strong AI,’ which is AI capable of performing any intellectual task that a human can, which, you know, those are vaguely terrifying classifications.
Matt: Yes, I would agree.
Lauren: The weak and the strong AI. But I did think that they were pretty great catch-alls for just kind of defining what that is. So yeah, if you're not familiar, if you weren't familiar before we jumped into this episode without even defining AI as meaning artificial intelligence, that's how ChatGPT would define itself.
Matt: Huh, good to know. I mean, and at least it was pretty accurate on that one.
Lauren: Yeah.
Matt: The one thing to point out too – for those of you that have not spent a great deal of time or even a little bit of time in ChatGPT or Bard or some of the others – is that oftentimes they're wrong.
Lauren: Yes.
Matt: So, you know, if you think you're going to use it for a search engine and replace Google with it right now, you should think again, that is definitely not something you want to do right now. It's definitely getting better, but it's nice to see that when Lauren asked ChatGPT to define or describe itself, it was pretty accurate. But there are many other things that we've asked it in the past for fun, or we've heard some of the horror stories where people have had it – or there was actually the one where the attorney used it to write his closing arguments.
Lauren: No.
Matt: And the jury came back and voted guilty on his client. And then the guy who was found guilty countersued his attorney for using ChatGPT to write his closing arguments. It had actually cited a court case in his closing arguments that was a fictitious court case and it was used to sort of set precedents. But it turns out that court case didn't even exist. ChatGPT just made it up.
Lauren: I wish y'all could see my face right now. I am genuinely gobsmacked.
Matt: Yeah, it's amazing. So yeah, please understand this is a great tool and this technology is great, but it's got a long way to go. It's nice that it gave you a pretty accurate sort of description of what it is and who it is.
Lauren: I think that's like an overarching point of caution that will get repeated multiple times throughout this episode, is that you should never take it at face value.
Matt: Yeah.
Lauren: Like it's always something whether you need to edit or fact check or whatever. It's a tool. It's not the end-all-be-all –
Matt: Yeah.
Lauren: – Of content creation or defining anything.
Matt: Yeah, and if any of you out there ever listen to our good friend Brian Fanzo, who talks a lot about NFTs and Web3 technology, one of the things he always says – and he always puts on all of his show notes and everything else – is DYODR, which is basically do your own damn research. Always, always, always do your own research.
Lauren: Absolutely, that’s great advice.
Matt: Especially when you're talking about things as highly technological as AI or, in Brian Fanzo’s case, NFTs and Web3 technology, and stuff like that, so…
Lauren: Advice to live by, honestly.
Matt: Agreed.
Lauren: So then in that case, you do my research for me now. Matt, you have had a little bit more experience exploring AI and stuff like that. You've been to a couple of conferences that the main focus of the conference – I think we've all been to conferences this year that have had at least a couple of sessions on AI – but you've been to some that are AI-focused specifically and exclusively. So do you have a better definition, or a different definition than what ChatGPT has provided me?
Matt: I don't know that I have a better definition. I did attend several this year, and I think the most beneficial one I attended was Marketing AI Con, and it's the one that's put on by the Marketing AI Institute and Paul Roetzer and his team. And that was a great, great event. And there was a lot of good information there. But what I will say is – to add to your definition or actually ChatGPT's definition of itself – the other thing that you may hear often in the same sentence as AI, or ChatGPT, or generative AI, is LLM, or Large Language Models. And I would just add that basically that's how these tools are trained. And that's the underlying technology behind a generative AI tool like ChatGPT. And so an LLM, what'll happen is people will feed it stuff from the internet, stuff from textbooks, articles, other sources. And that's where ChatGPT gets all of its information from. It's from this LLM and that's how it learns. And that model learns from that text data and it'll identify patterns and relationships. Unlike something like Google search, for example, it's not like ChatGPT is a query tool where you can type something in, and then it's immediately scouring and indexing the global internet and everything else. It's actually pulling from a database that it already has at that LLM. Yeah, I don't know if I make your definition any better, but I can't add context to how these tools kind of work and how they take on that human-like persona and it's through these LLMs and the way that they're trained and fed data.
Lauren: I mean, for me personally, I thought that was a nice addition. I'm also very impressed that you were able to say LLM consistently, correctly, every time, because even in me just saying it right now, I was so tempted to say MLM. So.
Matt: We definitely don't want to talk about MLMs. Those are bad, those are real bad.
Lauren: No, that's not this podcast.
Matt: Yeah. Let's not get into that.
Lauren: No.
Matt: Nope.
[11:28]
Lauren: Let's talk a little bit about how creators can use AI. I think it's something that a lot of us are aware of right now, whether or not we're using it. I think people are using it to different degrees, maybe cautiously, maybe incautiously. Maybe some people should be using it a little bit more cautiously. But in general, we have this relatively new tool at our fingertips. How can we use it?
Matt: I think the main way, right now, that most people are using it with any sort of success rate and efficiency rate is using it to buy back time in your day. We all know that time is the one thing that you can't get more of. You can make more money. You can get more friends. You can make more products, but you can't get time back in the day that you lost. Any small business owner, creator, author, the goal is to be able to maximize everything that you do in a given day towards the betterment of your business or what you're doing or your brand. And so if you have tools at your disposal that are powered by AI, that can help you greatly speed up the process of a lot of the work that you do that may be very manual in nature right now, but could be completely automated in the very near future, if not now. I think that's a huge, huge win.
Lauren: Oh, absolutely. I think that's the number one goal –
Matt: Yeah.
Lauren: – For any kind of creative person is to, like you said, find a way to streamline your day so that you can devote more time to the, either creative aspects of your job, or your business, or your passion, or the things that only you can uniquely bring to the table.
Matt: Yep.
[12:49]
Lauren: So what's one of the things that AI can do, or help you with?
Matt: Yeah, so things that AI can help you with that will save time in your day, your week, your month, and especially some of the things we use it for, or creators that we talk to use it for. One of the things is research. So a lot of what a creator does, whether you're writing fiction, nonfiction, whether you're writing up a technical brief for a product, it doesn't matter. You're doing some form of research on a fairly regular basis – and research could just be a fancy word for you're reading a bunch of articles every day to make sure you're staying current on the particular topic that you are writing about or wanna write about.
Lauren: That's research.
Matt: Yeah, absolutely.
Lauren: That's what we call productive procrastination.
Matt: Okay, yeah, absolutely. However, at the end of all that research that you're doing and all that gathering of data, you've got to somehow take that and turn it into whatever it is you started out to do, a summary for the stakeholders at your company or an article that you're going to publish on your blog next week or any number of things that you're wanting to create from all of that data and research that you just poured through. AI does not actually replace doing the research – or it shouldn't, not right now, as we talked about. It is not a replacement for Google search or any of those other things. You will get back some false information, hands down. So don't use it to replace the actual act of doing research. And if you do so, be careful, and make sure you go through with a fine tooth comb. But it is a great way to maybe aggregate a lot of that data or, even better, summarize it for you. So instead of going to ChatGPT or Bard to one of the others and saying, hey, can you find me five great articles that talk about the new way Taylor Swift is gonna sell tickets to her concerts?
Lauren: Please.
Matt: It probably won't find five great articles for you. It might, but chances are it won't. What you should do is find those articles yourself. That's a simple Google search, but then use ChatGPT to help you summarize the important points of those articles. That saves you a lot of time. So where you might have to spend hours reading all five of those articles, and then extracting out some of the key points that you think will help further your argument or the article you want to write, dump those articles into ChatGPT and have ChatGPT read them, and summarize them for you, and pull out what it thinks are the best points to make.
Lauren: What was the prompt that you – you mentioned this earlier today when we were talking about it, actually, and there was a, you said something about, like, ‘ask it to explain it like it was explaining it to a class full of high school students?’
Matt: Yeah, absolutely.
Lauren: Or something like that, which, that sounds great, honestly, when you phrase it like that. It's like, yeah, explain it to me like I'm dumb.
Matt: Yeah, 100%. If it could understand me prompting it to explain it like you're explaining it to my mom on iPhone, I would use that, but I don't think ChatGPT would get that.
Lauren: Not yet, maybe someday.
Matt: Yeah, absolutely. Like, asking ChatGPT ‘can you break this down and explain it in a way that a room full of high school students, or take it deeper and say a room full of fifth graders, would understand,’ you'd be amazed at what you get back. It's really helpful in that way, that's another way you can use it to help summarize information. Especially if you're dealing with a couple of articles you just dumped in, or some subsets of data and you're saying ‘hey, can you give me a pretty good analysis of this data or synopsis of these articles I just put in there, but also structure it in such a way where a room full of high schoolers would totally understand.’ That's what it's great for. It's not great for doing the research, right now. It's not great for some of those other activities. What it's great at is, once you've done that, saving you the time of having to read all that stuff, and then do a summary, and then extract out all the key points, and then make those work in the frame and the context of the article, or whatever it is you're doing. So yeah, a hundred percent.
[16:26]
Lauren: I mean, all of that sounds great. And I'm definitely going to be stealing that prompt for the future. I do want to make sure that when we're talking about doing this, we're talking about doing this for personal use. This is not for content. I don't want to say like, go ask ChapGPT to summarize five articles for you, and then you're going to use that summary as a blog post for your own blog, because that is just stealing other people's content.
Matt: Yeah, I think that's one of those great areas right now in the whole sort of debate around generative AI and how you use that content. Nonetheless, whatever side of the fence you fall on, for the most part, I think you’re right, I think what we’re talking about here is using AI to help spark content ideas, creativity, to get you a jumpstart on what you're doing. So in the example I just gave where I said, have it take these four or five articles, give you a summary of it. I personally would not take that summary word for word from ChatGPT and then publish that. I personally would take that summary and I would make that my own. I would use that as my starting point. Or if you're writing an article, that's just probably one part of the article, one of the discussion topics. So take that and then just grow that from a seedling to an actual piece of full content. So I do think there are people out there that would argue with you that, hey, if it spits out a quality piece of content that literally did not plagiarize any of the articles and it truly is just a summary, publish it, but say, ‘hey, I had ChatGPT summarize these articles and this is what it said.’ I personally don't have a problem with that.
Lauren: I...No, okay, I'm going to back up here for a second. I do see your point on that, but I think my bigger concern with that is that I would caution anybody from taking any content that came out of ChatGPT and publishing it directly.
Matt: Of course.
Lauren: You need to edit it yourself. You need to read through it with a fine tooth comb. You need to make sure that it makes sense.
Matt: Says the editor.
Lauren: Well, yes. But yeah, you know, one of the things that you mentioned earlier was that it can be wrong sometimes.
Matt: Oh, absolutely.
Lauren: If you are using ChatGPT to draft content for you and then you don't go ahead and at least like, review that content before publishing it and it's wrong, no one's gonna say like, ‘oh, he was probably using ChatGPT. It's not his fault that he got it wrong.’
Matt: Well yeah, but nobody said don't review it, I think –
Lauren: But that's gonna be the caveat. That's the middle ground that I'll concede with you on. I don't feel great about just publishing that directly, but if you are going to, make sure that you are at least reviewing whatever it gave you. Like you're not just copying it directly out of the ChatGPT chat and into your blog.
Matt: I feel like you just gave out one of those disclaimers like you find on Tide pods that says do not eat these. I'm not going to call you Captain Obvious, but she's right. Yeah. I mean, anything that any tool spits out, somebody should be putting eyes on that and reviewing it for sure. I agree a hundred percent.
Lauren: Listen, it sounds like Captain Obvious, but I'm also literally in a Facebook group that is just people posting screenshots of ads that were clearly written by ChatGPT and not reviewed by anybody first, so it happens.
Matt: Yeah but again, you’re talking about somebody who probably has to generate a hundred ads a day for some agency, they hate their life, but that's my point. If I gave you a task to take these five articles on what a short stock is. You're already looking at me like – right? So here's my point. And so you're like, well, I don't want to read these five articles on what a short stock is. I don't even care. And I'm asking you to give me a one paragraph summary on them. You dump those five articles into ChatGPT, ask for a one paragraph summary on those five articles about what a short stock is, it spits something out. You review it. It looks pretty good. Is there really any need for you to make any changes to that, or can you just hand it to me and say, ‘hey, I use ChatGPT, I dump those articles in there. Here's what it gave me.’ What do you think?
Matt: Is that okay?
Lauren: I think that's okay,
Matt: Boom. Boom.
Lauren: But the key phrase in there is it looks pretty good.
Matt: Well, yeah.
Lauren: I still took the time to review it and it looks good. That's all I'm saying.
Matt: Nobody was arguing that.
Lauren: Just take that time.
Matt: Yeah. Nobody argued that point.
Lauren: Fine.
Matt: Captain Obvious.
Lauren: All right.
Matt: Okay. Let's move on.
[20:10]
Lauren: Okay. Do you want to talk about using – I don't even know if this necessarily qualifies as researching, but I threw it in here anyway – using ChatGPT to help you write SEO-friendly copy? I wrote SEO optimized copy, which I'm realizing as I'm looking at it, I think is the equivalent of saying the ATM machine or the ISBN number.
Matt: That's right.
Lauren: So I will be rejecting that phrasing. But you know, if you are someone like me who works in marketing or content creation in some way and phrases like search engine optimization give you the heebie jeebies, and you don't even know where to start with that, this could be a very helpful tool to help you take your content to the next level with SEO without really having to deep dive into it.
Matt: I agree. And for people like me who like SEO, and it doesn't give you the heebie jeebies, and you enjoy keyword research and some of those other activities that come along with it, it's still a great tool. Now there are other tools that can do that for you. And some employ AI, and some don't, and I’m sure those will be integrating it sometime soon but that is a great use of it sure and we've done that before here at Lulu.
Lauren: Literally just while Matt was saying that I was DMing our content team manager, who is also our blog writer, asking him what the name of the tool that we use is. Cause we do actually have a Google Docs plugin. That's the Semrush Writing Assistant, that is AI, that is helping us optimize our blog posts.
Matt: Yep.
Lauren: By making sure we hit certain keywords and phrases and stuff like that. It's a great way to gamify the experience of making your blog posts optimized for search. That helps, if you're like me and you're not a big fan of SEO. Gamify it using an AI tool that will tell you how good you're doing based on whether or not you've used X number of phrases and words.
Matt: Yeah. Also for anybody out there that's listening, probably not, but in my age bracket, the term gamify is something that applies mostly to millennials.
Lauren: Oh.
Matt: So at my age, I don't need to be gamified or have something gamified for me to enjoy it. But I can absolutely see where you would take some enjoyment and some gamification around certain tasks that you might find less than pleasant.
Lauren: I have so many questions right now that I'm not going to derail our entire podcast with.
Matt: Yeah, we'll gloss right over that. In fact, why don't you just edit that in post?
[22:23]
Lauren: Maybe I will. We'll see. So, okay, let's move on to another great way that AI can help. And this is actually one of the ways that I use it the most often that I use ChatGPT, pretty much every day. And that is for outlining purposes, and also for taking that big old hammer to writer's block.
Matt: Yeah.
Lauren: Because there's nothing more overwhelming than just staring at a blank Google doc with that little cursor clicking at the top of it. And you're like, I got nothing.
Matt: I agree. And we talked before a little bit about our backgrounds and things like that. And I have a background – my college degree anyways – is in English lit and creative writing. And there was a point in time where I enjoyed writing and I did it all the time, a lot. And as I got older, I found it harder and harder to write, mainly to get started. And even now, or up until recently, at Lulu, I did not write a lot of content. Just because I had such a hard time getting started. Like I just couldn't focus. I felt like I had so many other things going on. And luckily I have a team of people like Lauren that could just do it for me. That's what I pay them to do and they do it really well. But –
Lauren: We try.
Matt: What I was going to say is, through the onset of tools like ChatGPT, I've been able to fall back into a habit of writing more and enjoying it, because I can use ChatGPT to get me started. Outlines to me are one of the hardest, most daunting things. And even if it's not something that requires an outline, sometimes I'll sit down to write on a topic that I'm very well versed in, that I am passionate about. But still, the idea of that first sentence getting scribbled down, or getting started with that first intro paragraph is just really tough for me for some reason. Once I get started, I'm usually pretty good. Using ChatGPT, I've been able to sort of get over that hump and get back into writing a little bit more. And I really appreciate that that tool does that for me.
Lauren: I do have a very similar background to Matt. Undergrad, I studied English lit. And then I went to grad school, I have an MFA in creative writing. And at the start of that journey, I was very much the kind of writer that…I had no outline and no plan and no direction. And I was just winging it. I would usually clear my schedule about 24 hours before I had a writing submission due, and I would just sit in front of my laptop for 23 hours straight and go like, ‘all right, what do you got, let's go,’ and just start typing like crazy. And by the end of that journey, I was so completely dedicated to the outline, and to the, like, you need to have some kind of clear path in front of you, you need to have some kind of direction need to know where you're going which I still now to this day find that incredibly helpful. But that doesn't mean that it's not equally daunting to come up with that outline, and so to have a tool like ChatGPT that can help you generate an outline, if you tell it ‘this is the topic that I'm working on, can you help me come up with an outline for it?’ That's a great way to get you started.
And even if, and this happens, I feel like this happens almost as often as I get a successful outline or a useful outline from ChatGPT, it'll give me something that I'll be like, well, that's wrong. I don't wanna talk about that. I don't agree with that. I'm just completely disregarding that. It still gives you a starting point because now I know what I don't wanna talk about. And when I see those ideas that are like, well, I don't wanna talk about this because it would make much more sense for me to talk about this, this, this, and this. And now there I go, I'm off writing my own outline. And even if ChatGPT didn't actually give me anything for the outline, I wind up writing for myself, it still got me started.
[25:38]
Matt: Yeah, yeah. I was just thinking about, as you were saying that, I remembered another example of how ChatGPT can be used. It actually may or may not fit in the section and it's actually unrelated to writing, but I just thought it was fun and some of the listeners might enjoy this, some might not. But when ChatGPT first came publicly available this year, my wife would take ChatGPT and what she would do is she would put into ChatGPT, ‘hey, listen, these are the meals I wanna cook this week. And she would list out spaghetti with meatballs, whatever, whatever, and then ask ChatGPT to give her a grocery list for those meals. And it would. Again this doesn't have a lot to do with writing necessarily but it is kind of like an outline right?
Lauren: It is, sure!
Matt: But it also just shows you the limitless possibilities into what you can use ChatGPT and these other tools like it. We keep saying ChatGPT, but there are others, Bard, and some others.
Lauren: Yeah.
Matt: But you know, there's just a lot of cool things that you can do. And for some people, an outline for an article, a grocery list, anything that is really kind of the genesis of an activity that needs to take place can sometimes be really hard. So you can find really cool ways to get around those activities, essentially have them done for you.
Lauren: How were those? Were they, like, actually accurate grocery lists?
Matt: Yeah, for the most part they were, they were actually pretty accurate.
Lauren: Well that's just good to know now.
[26:53]
Matt: What else would a writer like you do with it?
Lauren: What else would I do with it? Actually, so probably the second most common thing that I use ChatGPT for is editing and grammar checking, which can be incredibly helpful. You know, that's a big part of my day-to-day work at Lulu, reviewing other people's copy and content. And I second-guess myself constantly. It's been a long time since I've had the basic grammar education. And ChatGPT is great for just double checking myself with and just saying like, ‘you know, I think somebody used the wrong form of past here, but is this the correct use in the sentence?’ and copy and paste the sentence in there.
Matt: Right.
Lauren: And you know, it's not 100% perfect. It's not always right. But even that will help me then be like, if it gives me an answer and I say, I don't know if that's right. Sometimes that's all I need, confirm for myself that my instinct was correct and it was wrong.
Matt: Yeah.
Lauren: Then it's kind of fun a little bit to push back at ChatGPT and be like, are you sure about that? And then have it be like, oh, no, you're right. Yes, I am smarter than this generative AI software, another day at work.
Matt: What Lauren doesn't know is it's trained that anytime you push back, it concedes.
Lauren: Shh. Let me have this one.
Matt: All right, you can have it.
[28:02]
Lauren: Earlier when Matt was talking about using ChatGPT to summarize articles or other content and explain it back to you, try doing that with your own writing. If you're doing technical writing or if you're trying to explain a concept to an audience of people that might not be familiar with it in some way, when you're done writing it, throw it up into ChatGPT, or any other summarizer of your choice, and see what it spits back at you. And if it gives you something back that is inaccurate or doesn't properly explain what you're trying to say, then maybe try to explore where that disconnect is.
Matt: Yeah.
Lauren: And how you can change your writing to make it possible for ChatGPT to explain it back to you in the correct way.
Matt: Yeah, no, I think that's a good idea. So kind of using it as a low budget developmental editor in a way, I think that's pretty cool.
Lauren: Yeah, although, like we've been saying, and like I will continue to say, ChatGPT and any other AI is not a replacement for an actual real live editor. Please, if you are writing a book, hire an editor. Even if it's just one, you don't have to hire every single kind of editor, but please do not publish a book without having at least one other human being look at it.
Matt: I won't argue with that.
Lauren: Thanks. That's because you have two built-in editors on your team who will do it for free.
Matt: A minimum of two – not for free, by the way.
Lauren: That's – okay, that's true. I guess we're getting paid.
Matt: Yes. For everybody out there listening. They do get paid. I promise.
Lauren: Do I?
Matt: Well, we pay Lauren in Disney paraphernalia, but the other one gets paid in cash.
Lauren: Well, that is actually true.
Matt: It's probably news to you as well. But all right.
Lauren: It’s fine, I'm good with my deal.
[29:31]
Matt: Yeah. Let's talk about one of the other ways we definitely use it here at Lulu. And I know for a fact, a lot of other marketers that I talk to and peers also use it for these things. The simple act of using tools like Bard and ChatGPT for drafting small amounts of marketing copy, promotional copy, social media posts. You might need a little copy for a landing page you're creating to just capture email addresses at a particular event. Subject lines are my favorite thing to use it for. Everybody gets tired – if you do this often – everybody gets tired of trying to come up with, you know, clever, winning subject lines that will get people to open emails, and it's a great tool for that as well. There are some really cool new tools on the market that can help you draft social copy that can kind of match the tone of your brand. It will be keyword optimized, obviously, for the most part, and can use relevant hashtags or things like that. There's some really cool ways that you can use this stuff. And this will really carry you a long way with the algorithm gods that are overseeing all the social media channels.
Lauren: Always.
Matt: The Loki of TikTok, if you will, or whatever that might be.
Lauren: Don't you dare insult Loki like that. My God of Mischief is not having anything to do with Twitter.
Matt: But once again, if you use these tools for that, the caveat here is as we talked about, obviously review the copy, make sure that at least it's accurate. Or also try creating alternate versions of copy, you might have a piece of copy that you and your team are happy with or that you're happy with. Why not see what other versions might look like? Drop that in there and just say, hey, can you give me another version of this or something like that?
Lauren: Yeah, that could be a great way, if you're looking for ways to kind of level up your social media ads, almost all of the social platforms at this point will provide space for you to provide alternate copy for an ad. So if you're running a Facebook or an Instagram ad, and you have the main ad with the graphic and the copy, and then Facebook will give you the prompt of like, do you want to do I think up to five alternate versions of that where it's still the same ad, it's not a different ad –
Matt: Yeah.
Lauren: – But it'll give you different versions of copy. And AI would be great for that. And I think, actually, that Facebook Business Manager now has a built-in AI tool in their ad builder, where it'll give you alternate versions of the copy that you've already drafted and posted in there.
Matt: There's some similarities there too, for those who are in the world of paid search traffic and worked out of Google ads at some point, there was a point in time where they introduced dynamic ad structures, and what would happen is you would give it a couple of images, you would give it a couple of headlines, different pieces of text, you'd give it some other descriptors, lines of text, and then it would dynamically render an ad at the time somebody searched for something, pulling from those different pieces that you gave it based on what it thinks would resonate most with that searcher. I mean, that was a form of machine learning and AI right there. And this was years ago where they started doing that. And it's just gotten better and better over time for the most part. So I think that's another example very similar to the Facebook one you gave, where having those alternate pieces of copy, whether the tool generated them or not, but the ability for the tool to understand, ‘I'll use this headline with this descriptor and this image, because I think this searcher is gonna resonate with that the most,’ that combination is really cool.
Lauren: It is really cool and it's another great example of how AI is going to buy us all back some time.
Matt: Yeah, I mean that's the name of the game.
[32:50]
Lauren: Yeah, always, it always is. And then, also using it for titles. I will probably be using ChatGPT to help me draft titles for every episode of this podcast. I've got a list in here right now that were suggestions for this episode. I don't know if any of them are going to be verbatim exactly what the title winds up being. But even if it just gave me some ideas and I say, well, what if I take this element from this title and this element from this title and combine them and switch out one word? And it's a very helpful way to get started.
Matt: Yeah, and I think for people in the world of content creation, people like me and Lauren, for example, who might obsess over what a title of an article is going to be because we know how important that title actually is. I've literally before sat at my desk 30, 45 minutes trying to come up with a title. And if this tool can spit out ten in a matter of milliseconds, even if I don't use any one of those ten, I bet it gives me a leg up on where my thought process might go and help me come to a title faster on my own. Or I just say, ‘hey, those are great, but give me ten more. The tool is great for that, too. ‘
Lauren: Oh, it is, yeah.
Matt: If it spits out something that's heading in the right direction, tell it. Say you're on the right path here, but I’d love for you to explore the concept of this a little further, maybe give me ten more examples. And it'll do that. It takes directions really well. And so again, these titles that you have here, like you said, you might not use them, but something in one of these five titles might spark something in you. Or you can go back and say, ‘hey, those are great. But this one is really on the path to perfection. Can you give me five more like this one?’
Lauren: Yup.
Matt: Five more variations of this title.
Lauren: And it is great at that.
Matt: So saving minutes in a day, an hour in a day, all that stuff adds up and it all brings us to a level of efficiency that we've not seen prior to having our hands on this technology in such an easy way.
Lauren: I think a lot of this episode kind of sounds like it is more focused on content creation and nonfiction writers. If there are any fiction writers listening, I hope you're kind of drawing those dots for how this can be relevant to you too, because it absolutely is. This is a very helpful – any kind of AI tool is very helpful for all kinds of writing.
Matt: Yeah.
Lauren: But this was another example that I just thought of, of how I've used this is naming characters. I'm the worst at it. I've always said that's why I like writing fanfiction better than original fiction because the characters already have names. But ChatGPT, you go in there and you say, ‘hey, I need a name for a contemporary 20-something-year-old white male. And it'll give you 50 names. And even if you reject all 50 of them, it's a starting point.
Matt: Always go with Matt.
Lauren: No. I'm officially striking Matt from every draft of every – actually don't know if I've ever used the name Matt. Sorry.
Matt: Sucks to be you. Or your readers.
Lauren: What readers?
Matt: All five of them.
[35:26]
Matt: All right, so we've spent a pretty decent amount of time here talking about some of the ways you can use AI right now, as it is found in the wild, in these tools like ChatGPT and many others that are out there. And I mean, you can find some great articles out there just on the various different tools that will help you achieve certain things. Real quick, we will touch on some of the things we personally feel like you should not be doing with AI right now. And I'll get us started. This is going to be the one that Lauren obviously honed in on a little bit earlier. You should not be releasing AI-created or assisted content in the world without at least vetting it and potentially editing it.
Unlike Lauren, I will concede to the fact that there's a possibility you could get something from one of these tools that doesn't need any editing depending on what it is, but it still needs to be reviewed and vetted. You have to check it for accuracy. You have to check for context. You have to make sure that it makes sense for the audience that you're creating the content for.
Lauren: I think that's advice for any piece of content ever, regardless of who or what wrote it. You know, just give it a once over before you go anywhere with it. I stand by that advice and I will continue to bring it up at any available opportunity. Just review your stuff.
Matt: Yeah.
Lauren: Because people always find the typos, too. People always find the mistakes. The number of social comments that I would get, like the one time ever that I wrote like, ‘an’ instead of ‘and’ or something like that, I'd be like, okay, all right. Well, at least you know a human wrote it.
Matt: Oh, they're waiting for it.
Lauren: They are, there're sharks in the water and they're smelling that blood. So don't give them that opportunity.
[36:59]
Matt: That's right. Another thing you should not do with AI is have it write an entire book for you.
Lauren: Please don't.
Matt: And then plop that thing right out there onto a third party retail channel – which is happening, by the way. So be very careful when you buy books, especially from authors and creators that you don't know, or you've never heard of. Please do not use AI to write an entire book for you. And if you do, if you're not gonna take that advice, which is really just my advice or my opinion, and probably Lauren’s, at least for the love of whoever you pray to, follow what we've said and review it, vet it, edit it, make sure that it's not just another pile of garbage that's gonna enter into the stream of white noise out there on the retail channels.
Lauren: There is actually a really cool book on the Lulu Bookstore that I really like, that’s called Conversations With AI. And the creator very openly said, ‘I came up with this list of questions that are –’
Matt: Right.
Lauren: You know, questions like ‘what is the meaning of the universe’ and stuff like that, and asked ChatGPT these questions and then recorded the questions in a book. And I think also it's illustrated using, I think, Midjourney. They're not passing it off, the author or creator of this book is not passing this off as original content that ‘they wrote’ that ChatGPT wrote, that's the framing of the book. So we're not saying don't do that. That's really cool. It's a really cool project that I recommend you check it out, if that sounds interesting to you. But they're very upfront about the fact that this is a collaboration between a human and AI.
Matt: Yes. Publishing a book that's a conversation between you and ChatGPT, that's pretty cool. I agree 100%. I don't think there's any mistaking what's going on there. But again, what we're saying is don't use an AI tool, a generative AI tool, to write an entire book for you. And then you pass it off in the marketplace as something you created on your own. That's just gross.
Lauren: Right.
Matt: Don't do it.
Lauren: No, that's – don't do that.
Matt: And it's probably not going to be very good to be honest with you.
Lauren: No. No, it probably isn't. And I think that's where we all circle back to with this repeated mantra of, you know, make sure that you review any content before you publish it, anything like that. I know there are a lot of people that talk about the fear of ‘AI and generative AI is going to replace human creativity and it's going to replace copywriters or authors and I don't think that's true because as a person that has used ChatGPT to create content, I don't see it ever out-writing a human.
Matt: No, I agree, a hundred percent. And like we talked about at the beginning, we're talking about how to use these tools to make your life better, make your job easier, make what you do a little more efficient, all to elevate yourself as the creator of the content, not to necessarily highlight the tool.
Lauren: Right.
Matt: So a hundred percent.
[39:41]
Matt: And then, you know, lastly, again, this isn't an episode about the future of AI or anything like that, but is there any one last thing you might want to say, Lauren, around what the future of AI might look like or how it might affect what you do or anything like that?
Lauren: Not yet. There's definitely room for an episode entirely dedicated to what the future of AI looks like, because it's constantly evolving. We are recording this episode a little bit ahead of when it's going to get published. So, actually, at the time of the recording earlier this week, the Biden Harris administration passed an executive order on to the Department of Homeland Security about promoting the safe, secure, and trustworthy development and use of AI.
Matt: Right.
Lauren: So, you know, even between now when we're recording this episode and the time that it gets published, who knows how the landscape is going to have changed.
Matt: Yeah.
Lauren: So I think it's something that's really interesting to, like, look forward at the future of AI. And that's a topic that I'm excited to dive into later. But right now –
Matt: I think that's a good thing, though.
Lauren: Yeah.
Matt: So as we're already seeing, and what Lauren just alluded to, is really regulation. So unlike what we saw with the onset of Web3 technology and crypto and NFTs – which I think will come back in fashion – fight me if you think I'm wrong.
Lauren: Okay.
Matt: With AI, because of how fast and how impactful it's been, you're seeing stuff like this, like what Biden did, which is ‘we're gonna need some regulation, we're gonna need it pretty quick.’ So yeah, I agree with you there. As far as the future of AI, I agree with you. I don't know that there's any way to really do that, but I do strongly feel that it already has and will continue to change the landscape of publishing, specifically, self-publishing and things like that. I do think it'll change it for the better.
And then I think I'll just close this episode out, a quote that I took away from that marketing AI conference that I went to earlier this year. And please, whoever said this, forgive me for not remembering. It was either Paul Roetzer or somebody he was interviewing on stage, but what they said was basically, ‘this is the least capable form of AI that you'll ever use.’ That statement is very impactful to me because it basically says, for all it's worth, for whatever you're doing with it right now, whether that's not using it and being afraid of it, or whether that's starting to experiment with it and getting over your fears, or whether you consider yourself an expert in it right now – which is probably not accurate – this is still the dumbest version of it you'll ever use because it's changing every single day. So if you're finding value in it today, or you're gonna take some of these things we talked about back to your desk and try to make use of them, imagine if it's even the least bit helpful today, imagine how helpful it can be in your life six months from now, a year from now. I think more to come in future episodes when we have more time to talk about some of the more granular aspects of it.
Lauren: Yeah. And you know, in the meantime, if you're not currently using AI, or you've been on the fence about using it, I hope that we said something in this episode – anything in this episode – that convinced you to at least try it. And you know, I definitely agree with Matt that it's the future, you know, whether we're afraid of it or not. So I would say dive into it, explore it now, get familiar with it, get comfortable with it, see how it can make your day-to-day work, life, whatever the case may be a little bit easier. Yeah, let us know if you are using AI in creative ways that we didn't consider in this episode. Would love to talk more about it. And thanks for listening.
Matt: Thank you.