Trust and AI: Safeguarding Authentic Communication

AI: Your ally or adversary in communication? 🎙️ Lauren Sergy joins me on The Water Trough to discuss the nuanced role of AI in small business communication. Tune in for key insights and advice! #AI #SmallBusiness #Effective Communication

Transcript

Ed Drozda 0:10

Welcome to The Water Trough where we can't make you drink, but we will make you think. My name is Ed Drozda, The Small Business Doctor, and I'm really excited you chose to join me here as we discuss topics that are important for small business folks just like you. If you're looking for ideas, inspiration, and possibility, you've come to the right place. Join us as we take steps to help you create the healthy business that you've always wanted. Welcome back to The Water Trough folks. This is Ed Drozda, The Small Business Doctor, and today I'm joined by Lauren Sergey. Lauren is a communications expert. She has helped leaders across Canada, the US, and the UK to leverage the art of communication and public speaking. She's written two books, The Handy Communication Answer Book, named the best reference of 2017 by Library Journal and UNMUTE! How to Master Virtual Meetings and Reclaim Your Sanity. You can find more about Lauren's work at laurensergy.com. Lauren, welcome.

Lauren Sergy 1:27

Thank you for having me, Ed. I'm very pleased to be here.

Ed Drozda 1:29

And I'm excited to have you here. You always bring such energy to the conversation.

Lauren Sergy 1:34

I've got a lot of it to share.

Ed Drozda 1:36

And I appreciate that very much. So we're gonna talk today about the impact of AI in communication for small business owners. Oh my gosh. Let's just, let's dig right into that one.

Lauren Sergy 1:51

Oh, it's a meaty subject.

Ed Drozda 1:53

Oh my heavens, is it a meaty subject. I wanna start by posing this question. Is it possible in any way, shape, or form for AI to enhance communication as we know it?

Lauren Sergy 2:09

Yes, absolutely, absolutely. Despite all the drum and thrum over AI and certainly all of the anxiety that I feel over it, all the concerns that I have, it's a really useful tool. I use it quite a bit because it can enhance what we do. What it does not do is replace skill. And that's where my big concern comes in. Writing is hard. Communication is hard. Speaking is really hard. So, many people are keen to offload the hard work onto AI. Let it come up with the things instead of using AI in the way I think it's most useful, which is essentially a thinking partner or a second set of eyeballs. If you treat it like that, it can enhance the work that you did, but you must create the communication yourself.

Ed Drozda 3:06

It seems you're referring to communication as a vehicle. So the written word, something that is digestible, but can be transported. Are you referring also to the verbal form of communication here?

Lauren Sergy 3:21

Yes, absolutely. To me, they go together because very often if we're preparing to speak to someone, most people prepare by writing. And it's not that they're the same process. They actually take place in completely different areas of the brain. But those two forms of communication tend to bounce off of one another. And very often if you can get skilled at one, your skill in the other will also improve.

Ed Drozda 3:50

Interesting. When I think of AI and writing for example, I believe that my writing has to be of me. A lot of people write using AI, and when I say using I mean just that, and at least for the time being it seems to be apparent that it's not real. It's not human-based because it lacks soul. At least that's how I interpret it. As AI gets better, it will be better equipped to perhaps, hmm, have a soul? Am I barking up the wrong tree by sticking to my guns about the idea of owning my creativity? Am I missing something by not letting AI take a greater part in that process?

Lauren Sergy 4:41

I don't think you're barking up the wrong tree. What we don't want to do is make the mistake of ignoring what can be a very useful tool if used correctly for the sake of a certain creative ideology, which would be it must be mine, there can be no other interference, et cetera, et cetera. I am very concerned over the use of AI to replace the original thought of the human, and then the human simply glances over it afterwards and sends it out as if it's done. I am very concerned about the use of AI to do that because it has been used that way for quite a while. Written AI slop has been available on Amazon for ages. Especially infiltrating the children's book market. That was a really popular scam a couple of years ago. So it's a serious worry, but by the same token, disregarding the use of it altogether would be like saying, no I refuse to ride a bicycle instead of walking, because that's cheating. The wheel is doing all of the work. Well, no, you're still pedaling it. You're still doing it. It just, it can be more efficient. So certain tasks I find are really well suited for it and don't bother people overly much. I don't mind the use of predictive text in some ways. You know how in most of our email now, your email agent will suggest responses. Someone sends you a meeting request and at the bottom it'll have a suggested response of, thanks I got it. The meeting looks good. I've accepted it. That's fine. That is basically what you would've said. It's a couple of sentences. It's already typed out, read it, hit enter. You're good to go. But if the email that you're responding to contains any kind of detail and you know that they want to know that your brain is involved, those autofill responses can seem very tempting, especially to a stressed small business owner who is juggling a million things and doesn't have a lot of time. At first glance they look conversational. They might look like you've written them, but with a little bit of not even digging, with a little bit of an extra glance you might notice that there's something wrong. The greeting isn't the way you would normally greet someone or that you would normally sign off. But hey it's fast and you didn't look over it quickly, so you hit enter, you send, that's when it starts to become really problematic. The person reading it will pick up on those errors. They will pick up on the weirdness of your sign off, and a little bit of trust erodes. It's that little bit of trust erodes, and then you do it again and it erodes again, and it becomes a bit of a snowball. So, yes use it. It is very good in some applications, but it cannot replace your own actual effort completely. And I'm worried that that's what, I know that that's what people are doing because I deal with the outcomes of it all the time.

Ed Drozda 7:48

Erosion of trust. I use the term it lacks soul. But I think lacking soul is the disease. But the symptom is the most important part in this case, and that is the erosion of trust.

Lauren Sergy 8:03

Yes, and trust is everything in communication. It's everything because if the person doesn't trust that the information you're giving is good, that it's you who's saying it, that you cared about the response, they stop listening, very quickly. Or they start doubting everything you say. Now you have to prove that what you're saying is trustworthy. It adds to the workload of your communication. I'm gonna butcher this quote, but there's a really great quote from Stephen Covey, which is, oh let me see if I can get this straight. When the trust account is high, communication is quick and easy, or it's something to that effect. That absolutely hits the nail on the head. When people trust you, you don't need to work hard to communicate with them. I'm not saying it's zero effort, but you don't need to work hard to convince them. On the other hand, if they start sniffing that it's not really you making that response, they get mad fast. You're seeing this in a lot of marketing communication, in a lot of social media communication, and that's one for small business owners. If you do use social media to communicate with your client base, that's a really easy stumbling block because AI can be very helpful with some of your social media content, but when you start using it to create posts whole cloth, long-written articles, people can tell when it's not your voice. It might not be obvious on the surface, but there'll be this little tickle in the back of your brain that the soul isn't there. What typically they're picking up on is that it's generic. It feels very generic, like anyone could have written it and nobody wrote it. But it has the shape of you. That's it. Just the shape and it's the uncanny valley effect. The whole notion of the uncanny valley effect is that if you get a robot, you get an Android that is human ish, human enough to make you say, yeah that's human, but there's something a little off. Maybe it's the way the eyes look at you. Maybe it's something with how the mouth moves or the cadence of speech and it makes you feel really creeped out. That is happening in business' social media posts. Your audience is not going to like that. You will lose your trust bank. You will lose it with the emails, you will lose it with the reports that you clearly didn't read. And I'm kind of going on about this a little bit, but one of my friends the other day made a really excellent point. She was reading through a whole pile of grant applications, and aside from the fact that many of them hadn't bothered to erase some of the prompts that they had clearly used... That mistake happens all the time. And dear listener, part of the reason I'm laughing is because Ed's jaw just hit the floor. They hadn't bothered to erase their prompts or to erase things like Claude writing back, would you like me to look up information on X, Y, Z at the end of it? Aside from that many of them, she said you could simply tell that it was not the person writing it because it was so generic. It was giving the correct answers to the grant application questions, but there was no meat behind them, and that's the trap that we fall into with AI. There's no meat behind what we're saying and people pick up on it and they pick up on it faster than you'd like to think.

Ed Drozda 11:43

I can definitely relate to this. I've not seen the prompts showing up, uh, or the responses like that coming out. I believe that there are those who don't go back and review, and this is what they get.

Lauren Sergy 11:59

Yeah, and there's good reason why this happens with AI too. You have to think about how large language models work. They are predictive text. The way that it works is it has a massive amount of whatever platform you're using is trained on an inconceivably large amount of data. And what it looks for is the next most likely word to whatever it is you've said, and whatever it is you want it to say. Which means it's always going for the most common word, the most common phrase, and that's what it will always give you. That's generic.

Ed Drozda 12:38

Very much so.

Lauren Sergy 12:39

So if you go in and say hey ChatGPT, I'm giving a talk on basket weaving. It's 20 minutes long, write me a 20 minute long talk on basket weaving with these points in it. It is going to spit out something that sounds like everybody's Ted Talk, with the same pauses, the same beats, the same rhythm, but with a few points of basket weaving thrown in for good measure. Cause that's how it works. That's why we know, that's why we know it's not you.

Ed Drozda 13:17

You made the point that small business people are often juggling so many different things, and that any tool that would help them reduce that load is gonna draw their attention. Mm-hmm. AI by default is going to attract small business people, so how do they integrate this effectively and not erode that trust

Lauren Sergy 13:41

Here's one of the interesting truths about it. It's gonna seem very counterintuitive. Mm-hmm. If you want to use AI to help you write your emails, to help you write your reports, do the posts, et cetera, you actually have to use it a lot. Whatever platform you choose to use, use it as a thinking partner frequently. Frequently. If you're pondering a question and your brain is a little bit stuck. Open it up and have a little conversation with it, which is basically a conversation with yourself. Go back and forth because that way what the AI is going to do, and you will need, if you wanna do this, you are going to need to have an account on one of these programs so that it can save your data. It will learn what your voice is. It will learn what your most likely next word is. It will start to pick up on your patterns, but it needs that data from you. So if you have a lot of conversations on it, it will start to be able to better emulate your voice when you say okay, create this report for me. With some AI platforms you can get an account that does not share its data around, walled garden formats. At least that's what they say. Yeah, user beware. You can do this in a way that keeps your data within the confines of your own account. So if it reads what you wrote, if it has conversations with you, it learns your voice and then it gets better at imitating it. At that point when you use it to write something your review gets easier over time. The other big thing is do not treat the AI stuff ever as the final copy. You still want to, as I call it, touch. You still want to touch every word that's in there. You might not need to touch them as long. It's gonna be quicker, but you still want to have contact with every word that's in there. A recommendation I give people to make sure that it sounds like you is to read it out loud. As soon as you read it out loud you will know whether it's your voice or ChatGPT's voice. You'll pick up on that right away. After a few rounds of this you'll become more sensitive at picking that up in written form. It's like any skill, you get faster at it the more you do. Using it frequently really does make a big difference. And then deciding which activities give you the best bang for your buck when you're using AI. Because you might be thinking, oh the easiest use is to open up one of the chat bots and help me compose my stuff, when really the better use of it could be to set up a system within AI that automatically processes all of your receipts and spits out user reports for you at the end of the month based on whatever expense input is in it. That can chug away in the background while you do the harder communication work. So being very thoughtful about how you're using it and where it applies. Repetitive tasks, numbers-heavy tasks, those tend to be really well suited for AI use and AI automation. Then you can leave your brain open for the harder stuff, which is usually the communication.

Ed Drozda 17:01

That makes perfect sense. That is where I see it having a value. I do not feel the same about creativity, and most certainly don't see it as being a means of communication for me. Um, but yet it is going to be used that way. A replacement for what you and I call communication. Yes? It's gonna come to that point, at least for those who are so inclined.

Lauren Sergy 17:29

It already has for those who are so inclined.

Ed Drozda 17:32

Oh, okay. Tell me more because I think it's important to be aware that it's coming, what it looks like, and what that means.

Lauren Sergy 17:42

What it looks like are reports that are full of errors. I'm gonna be very blunt here. Reports that are full of errors, hallucinations that do not get caught. Communication that sounds generic and does not engender audience trust, and all of the other pitfalls that we've spoken about. We know that these are the pitfalls because they are already happening. That's the big thing. I do think that you are going to see more companies attempt to save on staffing labor, or attempt to increase employee workload by requiring them to use AI as a communication model. It does lead to burnout. It's so easy to put out that report really, really fast, so let's demand turnover for these reports faster because they can use AI to do it. I have seen that happening with my own eyes, and I do think it will eventually get to a point where most people realize, and it becomes best practice to no longer do that. Because wow, look at all of these case studies where it bit those companies in the backside. There's a number of case studies in various business news where companies are having to rehire. Laid off too many people thought AI could do it, and it can't. Situations where companies start to lose customer trust and that begins to negatively affect. These are already happening. So with any new technology there's always going to be the risk of overuse at first as we're figuring out what to do with it. But the pendulum, I think we'll swing back and we'll find the happy medium.

Ed Drozda 19:20

One of the things that's always on my mind about AI is the fact that we've long been required to think. I'm concerned that AI has the ability to reduce or even eliminate the need for thinking. As a small business person, I like to believe that everything I do has to go into one ear or the other, be processed by that thing in the middle there, and then action is taken. But if this tool has the ability to dumb me down, I don't know, that idea scares me. What's going on out there for the business world with AI?

Lauren Sergy 20:04

Thinking is a muscle. You need to work it out to get better at it. Now, just because you might lose the ability to think in one area doesn't mean you also lose the ability to think in another. If you learn your multiplication tables when you're in grade three and then never need to remember them or do that kind of counting in your head again because you have a calculator, it doesn't mean you don't know how to do math anymore. It just means that your thinking muscle in that area has atrophied a bit. So, when it comes to communication if you're allowing all of the thinking in the communication to happen through AI, because it's writing all of the stuff for you or whatever, you will lose the ability to think critically about the stuff that it's spitting out.'Cause you're not exercising that muscle enough. This is essentially a form of de-skilling. So you were a skilled writer. You were really good at writing emails before, but you haven't been doing it, and now you're not as good at it, and now it's a lot harder, because you've offloaded all of that onto AI. Can you learn it again? Absolutely. And you'll learn it again really, really fast. It's not the end of the world, but you do need to be aware that if you aren't actively thinking about the thing in which you have expertise, then you will lose some of that expertise through sheer atrophy. The morning of this recording I was watching a really interesting video from a software engineer. He was talking about the perils of using AI to code everything'cause AI is really good at coding. It's been noticed that the more that their program engineers relied on AI to create the code, the less they understood the code that was being written by AI. So they didn't know how to find bugs because they couldn't figure out how the code was written. They had lost that skill. The person who noticed this and put it on paper refers to this as cognitive debt. So now you have a thing that's being produced, but no one knows how it's actually produced. What happens when you run into trouble? No one can fix it. That's the cognitive debt, and that can happen with small businesses, with their communication. You sent out a whoopsie email. Shouldn't have sent that. How do I fix this? Probably don't know because you've been offloading the communication work onto AI, and now you're rusty, so you gotta build it up again. Yeah, and this isn't gonna happen with one or two uses. One of my favorite uses of AI is to help me edit things that I've written'cause if you've looked at your own work for too long, if you looked at anything you produce for too long, you stop seeing some of the mistakes. You can't see the forest for the trees. That's why editors are good. Well, we don't always have access to editors, especially at 12 o'clock at night when the due date for the thing is the next morning. So sure thing, that is a really good area where AI can step in, be that second pair of eyeballs. But the big thing is you still wrote it. You're still doing the thinking work. You just now have a thinking partner. That's not going to erode your skill. But offloading all of the work will, and it can happen across domains. If you are an accountant and you're now getting the program to do all of the accounting, you will lose your accounting skills because you're not thinking about it actively. So use the thing to help you out, but don't let it do all the work. You still gotta do some of the work.

Ed Drozda 23:50

So what I'm hearing you say is the task is assigned to AI, but oversight and processing of the task is ours.

Lauren Sergy 24:02

Yes, a hundred percent. In AI speak this is known as the human in the loop, and you never wanna get rid of the human in the loop. And that human needs to maintain their skills in that area. They have to do the hard thinking so that they don't lose their judgment, so that they don't lose their ability to think critically about whatever it is the machine has churned out.

Ed Drozda 24:28

Can I differentiate between the task and the critical thinking associated with the task? Let's say that I forgot how to build the widget. But I'm aware of what the widget must be. I know that it does or does not meet the objectives that I've set for it. Is it okay that I've lost sight of how to build the widget?

Lauren Sergy 24:57

I think that it's okay if you lose a little bit of that, but you need to retain enough of that knowledge to be able to find the errors or to be able to identify where there is an error. If you're part of a team, so you're a small business, you have 10 employees. You as the owner have a specific set of tasks. You are making high level judgements, high level decisions. Because you are in that position, that does not mean you need to be the person on the front line making the widget. You hire them because they know how to do that thing, but you know how to think about it from business sense. That's your task. That is your critical thinking ability. You could probably pick up if there's something that's gone wrong with the widget. And then say hey, what's going on here? What is this problem? Why has this happened? You are fully involved in this. With the notion of the human in the loop the human doesn't necessarily need to know how to put together all the code, I'm using code because it's a easy example. I do not code by the way, it terrifies me. That is not my skillset. The person who is doing the thinking about how the code is going to be rolled out doesn't necessarily need to be very skilled at writing all the little bits and bobs. That's why they have software engineers to do that work. But they should know enough to be able to notice if something's off and then go back to the software engineer and say, something's off. Can you find the error? And now the engineer can find the error. So you need to be able to think about these things at a fairly high level. Can you find if it's made a mistake? If not, there might be a problem.

Ed Drozda 26:46

If the code was misbehaving I would be able to see that it's misbehaving, but I wouldn't know why. So what I'm hearing is that because I don't know why I can't remedy the situation.

Lauren Sergy 27:07

If you know that it's behaving, but you don't know why, you know enough to ask questions, to find out why.

Ed Drozda 27:16

And I can inquire of the AI?

Lauren Sergy 27:20

You could inquire of the AI and if it's not giving you satisfactory answers, you could consult another expert and you can start finding those ways of learning what needs to be done to remedy it. The big thing is can you see that thing and say why did that happen, and be able to have the judgment to suss out or get a feel for whether the responses to that question are good or not. That's how at the high level you can work very effectively with with the work that AI might spit out. Not necessarily needing to know how to do every little bit in bob, but still being able to think critically about it. If you know enough to be able to dig and inquire then you can be pretty secure that the AI won't run away with things, or that you won't allow the AI to run away with things and do all of the heavy lifting for you. You'll be able to use it as a thinking partner. But you gotta be able to know when to say, uh, that seems off. Can I figure out why?

Ed Drozda 28:23

Keeping the human in the loop is being aware, observant, and being able to pick up on some very subtle circumstances.

Lauren Sergy 28:37

Yes, and depending on what the task is, you might need a specialist in that area to be the human in the loop. If you are running a software business, but you are not a software engineer, would be a really good idea to have a software engineer as the human in the loop'cause they're the ones who can do the better critical thinking. But you don't want them to just be like, I'm just gonna let the AI run all my stuff because they will lose their skills. So in a way, AI doesn't replace as much of our work as some may fear, or as some may hope. We gotta keep an eye on it and we have to keep asking it questions, and we have to do the quality control. We have to have our fingers in the AI pie, whatever pie it's producing. We wanna be touching all the ingredients.

Ed Drozda 29:30

Good visual girl. It's a little messy here. It's messy. It's mine was blueberry and I'm stained for life.

Lauren Sergy 29:40

Work is messy Ed, blueberries everywhere.

Ed Drozda 29:45

I love it. Lauren, time flies when you're having a good time and I have really enjoyed my time with you today. Before we part company, is there anything you'd like to leave us with?

Lauren Sergy 29:57

Don't be afraid of the technology. Don't be afraid of it. It is a tool. It's a pen. It's a wheel. Is it a very high powered one? Yes. But if you get to know it, if you use it, if you learn how to apply it with an air of curiosity, some patience, not for a view of how can this replace stuff I do, but how can this enhance stuff I do? Then you can use it as a very, very helpful partner in your work. And I think that is especially true of small businesses and small business owners. Don't be afraid, but don't let it do absolutely everything itself. Be the human in the loop and you'll be able to use it just fine.

Ed Drozda 30:45

Be the human in the loop. Lauren, thank you so much.

Lauren Sergy 30:51

Thank you, Ed. It was a real privilege.

Ed Drozda 30:54

Folks, this is Ed Drozda The Small Business Doctor, and here at The Water Trough I wanna wish you a healthy business, and also a reminder that AI is a tool. It's a tool that does not negate your responsibility, but it certainly does have the potential to enhance, improve, and change, yourself and your business. Until next time, be well.