Dr. Taylor Sullivan with Codility

Intro: [00:00:07] Coming to you live from the Business RadioX Studio in Woodstock, Georgia. This is fearless formula with Sharon Cline.

Sharon Cline: [00:00:17] And welcome to Fearless Formula. I’m Business RadioX, where we talk about the ups and downs of the business world and offer words of wisdom for business success. Today, I have a very, very smart person in my studio and I cannot wait to learn from her. She is the senior director of wait! Say it again. How did we say it? Product. Product. Okay, because there’s a different title. But your senior director of product. So I want to make sure I said it correctly at Codility. Yes. And welcome to the show, Dr. Taylor Sullivan.

Dr. Taylor Sullivan: [00:00:50] Thank you. I’m super excited to to chat with you.

Sharon Cline: [00:00:52] Me too. I’m sorry. I kind of butchered your title. It’s fine.

Dr. Taylor Sullivan: [00:00:55] Yeah, I butchered it. I don’t really know what I do a new title for.

Sharon Cline: [00:00:59] You, which is cool. Like you had a different title, and now this is a new one. Actually, it’s even next week. It’s brand new. So congratulations on your new title.

Dr. Taylor Sullivan: [00:01:06] Thank you very much.

Sharon Cline: [00:01:07] You’re welcome. One of the cool things about Taylor is that she is very, very well versed in AI, and that’s why I wanted her on the show, partly because it does affect my voiceover world, but it also affects pretty much every other aspect of the business world, even non business world. So serendipitously, I was released to the public yesterday, a year ago yesterday. So I kind of think it’s like, cool that you’re here on the one year anniversary. And we were just speaking before the show how you happen to be right in the epicenter when it did really become public. So can you talk a little bit about that?

Dr. Taylor Sullivan: [00:01:43] Yeah, I guess we’re kind of having a birthday party for it.

Sharon Cline: [00:01:47] Happy birthday I. Yeah.

Dr. Taylor Sullivan: [00:01:49] So I would say, you know, AI has been around for a long, long, long time. And specifically large language models have been around for a long time. Sometimes it’s referred to as generative AI. When ChatGPT dropped a year ago, that was the first time that it was really democratized it. It had a really accessible user interface, and it was easy to use. And so you didn’t have to be a programmer who is very skilled in, in writing code to, to prompt and to use these models. It had a slick new interface that anybody can use. So it just increased its accessibility so much. And that’s really what set it on fire.

Sharon Cline: [00:02:30] So AI has been around. So what I’m saying. Ai it’s different from what we’re talking about generally as ChatGPT.

Dr. Taylor Sullivan: [00:02:36] Right. Yes. Chatgpt is a is a subset of. So there’s AI. And then within AI there’s generative AI which is producing language not just consuming it, it’s actually producing novel content. Right. And then it’s based on a large language model. So the large language models are basically consuming practically the entire internet. And, and algorithms are operating on that body of content to essentially predict the next letter based on what is most likely to be the next letter, given its training data. So it’s not actually thinking or formulating real thoughts, it’s predicting one letter at a time, or which ends up being what’s the most likely next word. And so it’s it’s kind of just unfolding, but it’s really smart. So it unfolds in a way that sounds like natural human language.

Sharon Cline: [00:03:42] Right. So I actually looked up what chat because I’ve heard chat GPT a million times but it’s generative pre-trained transformer.

Dr. Taylor Sullivan: [00:03:49] Yeah. So that’s what GPT stands for. And GPT underlies a lot of AI tools. So chat GPT is kind of the interface. The way of accessing GPT and GPT has evolved over time. So when chat GPT came out, it was based on GPT 3.5, and it has since evolved and there is now GPT four. So if you go into chat, GPT, the free version is based on GPT 3.5, or you can pay 20 bucks a month, I think, and get the four.

Sharon Cline: [00:04:23] Wow. I wonder how different they are.

Dr. Taylor Sullivan: [00:04:25] They are different. Four is much more powerful. It can handle a lot more data and it can produce a lot larger pieces of data. I have not pulled the trigger to update to 3 to 4 because 3.5 has met my needs. However, I’m very tempted to give myself that for Christmas.

Sharon Cline: [00:04:43] Well, I imagine lots and lots of people pay $20 a month, including, I’m sure, businesses and.

Dr. Taylor Sullivan: [00:04:50] They actually are overwhelmed with demand from businesses to the point where they’ve had to kind of pause.

Sharon Cline: [00:04:55] Okay, so OpenAI, right? Is that what is the the very first chat, GPT sort of platform?

Dr. Taylor Sullivan: [00:05:02] Openai is the company that that that has been building out that that platform. Yes.

Sharon Cline: [00:05:08] So this is what’s so interesting about it is that it was just so quiet. It quietly kind of came onto the scene like. No. No big announcement. No big, I don’t know. Update of the cell phone of like, you know anything. It was no launch. Like what Microsoft does or anything. I mean, I was just thinking how interesting it is that it is kind of become so important in our everyday lives. But like a year ago, no one had heard of it. Like a year ago yesterday.

Dr. Taylor Sullivan: [00:05:35] Yeah. So interestingly, some people had heard about it. And so in in the tech world, people knew it was coming. So my company, for example, had beta access before it was released. We played around with it. We knew what was coming. I think we were still all surprised at just how powerful it was. But it’s it was not necessarily, you know, the dawning of something new. It was it was a natural evolution of, of development that had happened for a long time.

Sharon Cline: [00:06:09] So it was a business tool that was promoted to certain businesses, but not as much to the general public. So you knew it was coming, but the general public was kind of not part of that big.

Dr. Taylor Sullivan: [00:06:19] Correct.

Sharon Cline: [00:06:20] And then now it’s become so much part of the general public.

Dr. Taylor Sullivan: [00:06:23] Yeah. Well, I was I was saying earlier today, actually, to me, this is the most similar thing in our, in our history is when the internet came about. And I think that this has the potential to revolutionize things just like the internet. Did. You know the internet caught on like wildfire? There’s a learning curve. People learn how to use it and then they can use it while ChatGPT there isn’t even as steep of a learning curve. And so you can start to get value from it so easily and so quickly. And so they didn’t really have to promote it because it was spreading by word of mouth. And its value is so quickly obvious that, you know, it sells itself.

Sharon Cline: [00:07:06] It’s just interesting because obviously when the internet started to become so popular, there weren’t as many regulations around it because we were learning. Right. And it’s the same right now. Correct. What are you seeing that would need some more regulation?

Dr. Taylor Sullivan: [00:07:20] It’s a great question. And I will say this area in particular in particular is evolving super rapidly. There are several pieces of really big legislation that are kind of on the table right now. The EU has it’s called the AI act that has gone through, you know, a lot of deliberation and, and it’s it’s making its way through becoming a reality. It, it it’s, it’s a very interesting law because it, it puts the burden on the companies that are, that are providing these tools. But it it takes a consumer protection standpoint. So it’s going to be AI tools will be subject to things that you might expect from like product warning labels and things like that. So it’s a consumer protection versus in the US they’re taking more of of a potential for discrimination angle. So they’re like New York City has a law, local law 144. And if you use any tool that has an algorithmic process or has automated decision making, which kind of could mean anything if you, you know, if it’s adding two numbers to provide a score for a test or something, it could be susceptible to this, to this law. And you’re required to do a bias audit to make sure that the tool is not resulting in discriminatory outcomes for members of certain underrepresented groups.

Sharon Cline: [00:08:57] No way.

Dr. Taylor Sullivan: [00:08:58] Yeah. So that law has been interesting because it places a pretty heavy burden on the end user of these tools, you know, so my company has a platform that offers technical skill assessments for software engineers. So we’ve been thinking about this and we don’t currently use AI in our in our scoring or evaluation of humans. But you know, we still do some math based on how they how they answer things to produce a score. And so, you know, the implication of that law is if a customer in New York or a customer is hiring citizens of New York City, they may be required to put on their website the results of this bias audit.

Sharon Cline: [00:09:44] How is how, how are they able to determine what is going to encourage a discriminatory, discriminatory outcome? What is it that would what a name.

Dr. Taylor Sullivan: [00:09:58] Well, so in in like. There’s this concept of adverse impact in in the legal field. And that’s basically when a hiring ratio is different enough for one group than another group. That suggests there’s something fishy going on and there’s just. Discrimination or differential impact across groups using a tool, for instance, a hiring test. Okay. And so it’s just statistical ratios. And you’re comparing one ratio for this group to the ratio for this group. And if they differ by a certain amount that’s a flag.

Sharon Cline: [00:10:34] Got you.

Dr. Taylor Sullivan: [00:10:35] And like the EEOC and they govern all of this all of this area.

Sharon Cline: [00:10:43] Wow. So I’m sorry my I’m I’m listening. I’m not as I’m really like I can be smart in some things, but like this I’m learning. So I’m sorry if I ask questions that seem simplistic.

Dr. Taylor Sullivan: [00:10:55] No, no, I mean, it’s so interesting with with AI, what it has done is created this spectrum of issues and so on. The one end, you have folks that want to lock it down and regulate it. And and you know, it’s a little bit more of a, of a fearful standpoint. And many people are there, many companies are there. And then all the way to the other end of the spectrum where it it is viewed as a tool that really unlocks human potential. And, you know, in the right hands someone who can harness its power. It’s an incredibly amazing tool, and it can increase productivity. It can, it can, you know, automate some of the yuckier tasks that people don’t want to do and really kind of free up brain space for more complex, like human things.

Sharon Cline: [00:11:51] Right? So in the voice over world, obviously. Commercials have copywriters, right? So oftentimes I’ve heard that copywriting has been taken over by by. What do you what do you think of or ChatGPT. What do you think of this?

Speaker3: [00:12:09] Um.

Sharon Cline: [00:12:10] I’ve read some, and it’s really nuanced. It sounds great. Yeah, it looks great.

Dr. Taylor Sullivan: [00:12:17] Yeah. I mean, yes, it is very good at at certain things. It’s good at producing copy. It’s good at it’s good at brainstorming, giving you kind of initial. It’s almost like you have a writing assistant that does a first draft for you and then you as the expert, come through and, you know, tweak it, make it better. I heard someone say once, it’s like having a junior assistant that doesn’t know anything, but has read every book in the world.

Sharon Cline: [00:12:48] Oh my goodness.

Dr. Taylor Sullivan: [00:12:48] And remembers what they read, but they don’t actually know how the information goes together so they can give you facts.

Sharon Cline: [00:12:55] But but the human comes in to make it warmer. Yes.

Dr. Taylor Sullivan: [00:12:58] Well, and one other unique thing about AI is it does what they refer to as hallucinations. So it says things very confidently that just aren’t factually incorrect. So, so you you have to know, you have to be a critical consumer when you’re working with it. And so it can’t necessarily replace human writers because they will say things that aren’t true. Now as it evolves, we can expect to see that go down some. That’s kind of one of the like the hallmarks of a of a good MLM is that it doesn’t hallucinate as much.

Sharon Cline: [00:13:32] When you say Lm large.

Dr. Taylor Sullivan: [00:13:34] Language model.

Sharon Cline: [00:13:35] Gotcha.

Sharon Cline: [00:13:35] I’m going to start using that so I can sound smart. Okay.

Dr. Taylor Sullivan: [00:13:38] Yeah. Do it.

Sharon Cline: [00:13:39] I was reading an article, the Verve had an article that came out yesterday because it was the one year anniversary, and they were saying that since its initial launch nearly a year ago, ChatGPT has hit 100 million weekly active users.

Dr. Taylor Sullivan: [00:13:56] It’s the. It had the fastest and largest uptake of any app in the history of history. Wow. I heard though that threads did surpass it. Oh really? Initially, but I can’t. I don’t know if threads has kept pace.

Sharon Cline: [00:14:12] Right. Interesting. I had that for a minute and then I was like, I don’t like this. Okay, now listen, there’s only so many areas for my brain to go to be out there on social media. You know, it’s just a little much for me.

Dr. Taylor Sullivan: [00:14:25] I will say that ChatGPT is worth adding to your list of tools that you do use, though.

Sharon Cline: [00:14:32] That was a good plug. Well, you’re using it, obviously you use it. Tell me how you use it.

Dr. Taylor Sullivan: [00:14:38] I use it every day and. I would say the best way to learn how to use it is just to try doing stuff with it. It’s totally a trial and error process and you get better at using it with time. So the using ChatGPT sometimes is referred to as prompt engineering. So, you know, I’m an engineer and that I can write prompts that make the the tool do what I need it to do. So I use it all the time to just see if it will do stuff.

Sharon Cline: [00:15:11] So what do you do?

Dr. Taylor Sullivan: [00:15:13] So the one of the first uses I cannot stand meal planning. It’s just one area that I would be happy if I never did it again for the rest of my life. But I have two young kids. My husband and I both work full time, so it’s just hard to get meals, the groceries purchased for the meals plan the meals, make the meals. So I had it be a meal planner, and I asked it to give me three recipes that had five ingredients or less and take 30 minutes or less to make. Then take the ingredients for those recipes, put them on a grocery list, and put other common groceries with it. And it spit me out. A grocery list and three recipes.

Sharon Cline: [00:15:55] Did you make the recipes?

Dr. Taylor Sullivan: [00:15:57] No. I also don’t like to cook, which is the problem.

Sharon Cline: [00:16:01] Sorry. I just think it’s funny because you asked it to do something so specific and like, highly, you know, like important to you because you hate to do it, but you’re like, yeah, no.

Dr. Taylor Sullivan: [00:16:12] But at the same time, if it could cook the meals for me, then, then yes, I would.

Sharon Cline: [00:16:15] You’re almost there. Right?

Dr. Taylor Sullivan: [00:16:16] We’re almost there. Yeah.

Sharon Cline: [00:16:18] At the same time, though, if you did not ask it in the exact right way, will it not give you those things?

Dr. Taylor Sullivan: [00:16:23] It will. It will give you a version. But you can basically you can improve the quality in terms of how you write the prompt. And, you know, there are some basic guidelines that you can follow when writing a prompt. And I actually wrote these down so I would remember them.

Sharon Cline: [00:16:38] Nice. You did better than I did. Should have written down some things.

Dr. Taylor Sullivan: [00:16:43] And it was funny. I told Chad to give me ten examples of prompts that included these things.

Speaker3: [00:16:51] So you ChatGPT.

Sharon Cline: [00:16:52] What you’re going.

Speaker3: [00:16:53] To say about.

Dr. Taylor Sullivan: [00:16:53] Chatgpt? Yes, yes, of course, this moment. Okay. So the first thing that you’re supposed that that helps it is to give it a role, tell it who it is, because that kind of conjures its training set around. Like what profession profession am I in. And, and kind of generally speaking, where will the information be coming from. So you could say you are a meal planner and then you provide context. So what’s the nature of the task? You know, you’re trying to reduce the time that you need to spend every week on meal planning. So you’re looking for easy recipes. You’re looking for things that can you can make really quick. And then you provide it a very specific instruction. So, you know, like I told you, make give me three recipes, put them on the grocery list. Et cetera. And then and then it’s helpful to specify what format you want the outcome in. So I want a list I want a table. I want a picture Gpt4 can provide pictures. Gotcha. And you can say like factor you can factor in certain parameters. So like. Or you could, for instance, you could have it be a travel a travel agent and tell it, tell me three destinations the constraint and list them out. Describe them. The constraint is it should cost no more than $3,000 a person to go. And so by providing constraints you’re kind of narrowing what output it’s giving. And then it’s also super helpful if you provide examples. So here’s an example of what I’m looking for. And then do that ten times.

Speaker3: [00:18:30] Wow.

Dr. Taylor Sullivan: [00:18:32] Yeah, it’s pretty awesome. It it it writes amazing bedtime stories. So you can say, like, write a bedtime story using your child’s name and in it, like feature themes about X, y, z. And it should be about cowboys and it writes a beautiful story.

Sharon Cline: [00:18:54] Oh my brain. My brain is trying to catch up. So. I think this is wonderful because it is incredibly helpful. Like you said, I don’t have to come up with a story. I can use my brainpower for other things. Yeah. What other things am I going to be using my brain power for?

Dr. Taylor Sullivan: [00:19:13] So there are some things that it can sort of get at, but it’s much better done in the hands of a human. So. It can sort of solve problems, but what it can’t do is the initial decomposition of the problem and kind of planning how you’re going to solve the problem before you’re just generating solutions. It just kind of skips straight to generating solutions. And so as humans, what we can do is still when we are evaluating a problem, we’re really thinking through what what the aspects of the problem are. And you’re just engaging and interacting with the problem more than, than a tool like this would do. And so we’re still better at decomposing problems into smaller pieces and structuring kind of how we are going to go about solving the problem. How are we going to build the solution? What’s the logical order and what’s the logical steps that you should go through to solve the problem. And then as you’re executing, these tools are very helpful. But that initial problem decomposition, problem formation, we’re still really good at that. And so the more time we have to really think about the problem, you know, the tools can help with the solutioning, but they’re not as great at coming to know and understand the problem itself. And that’s a little abstract.

Speaker3: [00:20:43] No, no, I.

Sharon Cline: [00:20:43] Get what you’re saying. I mean, it is a little abstract, but but it makes sense because how would all of this seems abstract, frankly. Yeah. So. I’m thinking what you’re talking about is using brain cells in a different way. Like, I don’t need to come up with a creative story. I can just have. I can just read it because that’s what my kids want, right? They would just want a little story before they go to bed. But if there’s a really a genuinely deep problem that I need to spend time thinking about, I don’t have to worry about trying to come up with like, a little story. Let me just get these kids to bed and let me now really work on and use.

Speaker3: [00:21:16] Creative.

Sharon Cline: [00:21:18] Parts of my brain.

Dr. Taylor Sullivan: [00:21:19] Well, and, you know, in that example, what I did spend time on is thinking about what elements I wanted in the story that would resonate with my child. And I know that it doesn’t know those things. It doesn’t know how to combine those pieces of information and those pieces of the context to make a highly personalized story for my child. I know those things, and so I can tell it that it needs to have cowboys. I can tell it that it needs to feature somebody that goes potty in.

Speaker3: [00:21:47] The, in the.

Dr. Taylor Sullivan: [00:21:48] In the toilet. Yeah. But because that’s what he’s working on and I uniquely know that. And so I can combine these elements and formulate the problem that I give it to solve. But I’ve thought through how to how to formulate the problem.

Sharon Cline: [00:22:02] Let’s imagine that you told it you wanted that story, and then you’re like, this is a damn good story. I am now going to publish a book that has this story. How do you keep track of who is actually the author of anything?

Dr. Taylor Sullivan: [00:22:14] It’s a good, good question. And kind of one of these issues that I’ve just have decided that I’m not going to try to figure out the answer to. But it’s.

Speaker3: [00:22:21] Complicated.

Dr. Taylor Sullivan: [00:22:22] It is. And, you know, there’s certainly I would say there is much more friction in the AI generated image world because you can these these images are basically being constructed from every image in the world. And, you know, sometimes like a Getty Images logo will sneak into the image. So it’s like this is absolutely, you know, not not your own property.

Speaker3: [00:22:45] I was just.

Sharon Cline: [00:22:46] Reading that open AI’s best and most well-funded competitors mid-journey they’re in image generated AI company Pica, which I guess is also an image generated company, so I don’t even know. Actually, this is because I don’t live in this world very often, so I didn’t know that there was like a whole other subset like that. So they potentially could be taking images that really don’t belong to them and putting them all together, making a new image. But fundamentally, it’s not their material.

Dr. Taylor Sullivan: [00:23:18] That’s correct. And, you know, there’s a lot of outcry in creative communities because it is hard to regulate where the content came from and who’s whose intellectual property it is. So I think this is an area to watch. I think we’ll see some evolution going back to the to the legal side of things and not necessarily legal. But a few weeks ago, President Biden released an executive order. And one of the aspects included is that he’s kind of requiring these companies like OpenAI, to figure out, how are you watermarking content that’s generated by AI so that at least if you hand me a paper, I can see you use generative AI so I can as a consumer know, like maybe these are your ideas, but they might not.

Sharon Cline: [00:24:08] Be or imagine. Getty making it somewhat impossible for anything to be culled from the internet that way.

Speaker3: [00:24:16] Yeah. You know and.

Dr. Taylor Sullivan: [00:24:17] And that is happening too is it. Companies are making it harder to access their content and they’re embedding like watermarking. And so that it it would flag in these types of watermarking systems. So that’s happening to like the creative creative industries are reacting and are looking to protect their interests and are taking steps to do that.

Sharon Cline: [00:24:39] Doesn’t it feel like they need to be doing it 24 over seven because this is moving so fast?

Dr. Taylor Sullivan: [00:24:43] Yes, yes, I it’s it’s, you know, a horse race at this point.

Speaker3: [00:24:51] It’s so.

Sharon Cline: [00:24:52] Crazy.

Dr. Taylor Sullivan: [00:24:53] Yeah. It’s changing everything. So it’s again going back to the dawn of the internet like this is just as big. So this is the next big paradigm shift. And it’s kind of changing the whole world of computing.

Sharon Cline: [00:25:09] I was reading in the Verve article about how they said that this past year it’s just grown exponentially, but this next year it will grow even faster. Like just trying to keep up with it because it’s almost like feeding on itself. Somewhat.

Speaker3: [00:25:23] Yes, it is it.

Dr. Taylor Sullivan: [00:25:26] Well, sort of technically, but generally speaking, yes. Figuratively speaking, yes. It’s feeding on itself, but there’s also a lot of competition in the industry. And so it’s this race to have the best large language model. So you have big players here, you have Microsoft, you have Google, you have OpenAI, Apple. So you know, it’s it’s who can have the fastest, most accurate, fewest hallucinations, most powerful multi multimodal images text audio like that’s we’re going in in places that are probably going to be things we’ve never seen.

Sharon Cline: [00:26:04] Hmm. This is very powerful. Right. So where do you see this power being used for good? And where do you see it being used for evil? I don’t know if evil is the right word. Darkness. The dark. Because there’s always dark with anything. Light. I mean, immediately when the internet happened. Happened like the next. That was the next w-w-w, you know?

Dr. Taylor Sullivan: [00:26:26] Yeah. I mean, the potential for good is limitless. And I think, like I said before, it really will unlock human potential in a lot of ways. It’ll enable us to innovate faster. It’ll enable us to make people’s jobs more enriching and fun and engaging, because it’s going to automate some some of the yuckier tasks, like I said. And I think, you know, it will ultimately help humanity in ways that are quite, quite notable, making our lives easier, helping us think, helping us structure our thinking, helping us work through problems, that kind of thing. There’s also certainly potential for for wrongdoing. I think one of the best examples is deepfakes. You know, that’s kind of an alarming technology, and it can be used for very bad things. The probably the best use of it that I or the most salient use of it right now that I’m thinking through is, again, we we have a hiring platform and one of our products is a is a video interviewing tool where you’re assessing someone’s ability to code along with you and work through a problem. And, you know, there’s all these proctoring softwares out there that watch the person make sure that they’re doing what they say they’re doing. And and even that you, the, the job candidate, are actually the one they’re in the, in the chair. But with a deepfake, there is no way to detect that in a video interview if you’re talking to the person or not, because you are from every indicator. So that’s scary to me.

Sharon Cline: [00:28:18] So in real time, it could be as if if I were doing a zoom interview with you, someone could deepfake my image and voice. Yes, anyone could say something that I would. Even not say.

Speaker3: [00:28:32] Yes.

Dr. Taylor Sullivan: [00:28:33] And so the implications there are scary. You could have a world leader saying things that would incite a war.

Sharon Cline: [00:28:43] Do you think that that will encourage more of a premium on in-person activities speeches?

Dr. Taylor Sullivan: [00:28:52] It could. It very well could. I think right now the initial thing that it’s triggering is enhanced privacy. And, you know, your like restrictions on using people’s likenesses. And so I think we’ll first see legislative backlash and then we’ll go from there. But I think the legal world is trying as hard as it can to catch up right now.

Sharon Cline: [00:29:17] Do you see it being used in medicine? I’m thinking like the positive thing.

Dr. Taylor Sullivan: [00:29:21] It is being used in medicine.

Speaker3: [00:29:22] How is it.

Sharon Cline: [00:29:22] Being used in.

Speaker3: [00:29:23] Medicine?

Dr. Taylor Sullivan: [00:29:23] Well, it’s being used to help with diagnoses. So imagine you’re feeding it a list of symptoms, some descriptions of patients and using it to, you know, go through a medical book essentially and offer up what it what are the most likely underlying ailments. But then as a doctor you are knowledgeable and you can qualify or disqualify any of the things that it’s saying. So I think it’s being used as a as a reference. What’s really exciting is that it can put information together in new and different ways. So it also has the potential theoretically to come up with cures to things, because it’s it’s creating patterns and relationships and connections in ways that we might not have as humans. And so we might see it driving medical breakthroughs.

Sharon Cline: [00:30:21] It’s unbelievable.

Speaker3: [00:30:24] Yeah.

Sharon Cline: [00:30:25] Could I tell it? You’re a doctor, and these are my symptoms. That’s the kind of prompt I could give it.

Speaker3: [00:30:31] Yes.

Sharon Cline: [00:30:32] That’s so crazy.

Speaker3: [00:30:33] Yeah.

Dr. Taylor Sullivan: [00:30:33] It’s basically taking WebMD to the next.

Speaker3: [00:30:36] Level.

Dr. Taylor Sullivan: [00:30:37] Convincing us we all have Ebola.

Sharon Cline: [00:30:39] Every time I go on WebMD, the bottom line is, like, I’m not going to be here much longer. Like, it freaks me out, you know, because it’s always something significant. I’m sorry. I really feel like I’m not sounding intelligent at all because I am really struggling to keep up with what you’re describing. To me.

Speaker3: [00:30:56] It’s a.

Dr. Taylor Sullivan: [00:30:57] Lot. And, you know, I think about this stuff almost all day, every day right now in my role. So one thing that I didn’t really get into was, you know, how did I get into this? And like, what is my relationship to I? Well, I’m not an AI scientist and I don’t build AI systems. I am an I o psychologist, and I study how trends like this affect the skills that our workers need and how it’s going to affect their well being, their experience, their relationship with their customer, their their performance, their productivity. So I’ve spent some time studying like, how does this change workers experience, what skills they need, what skill, how to identify those skills, how to how to assess them, how to develop them. And so that’s more of the angle where I’m coming from here is basically how can we help prepare and optimize humans to collaborate with AI?

Sharon Cline: [00:31:59] What are some of the things that you’re finding as a trend that is affecting someone’s well-being, as they’re learning and using these models?

Dr. Taylor Sullivan: [00:32:10] I don’t I don’t know as much on on well-being, but there is one skill that is really showing up time and time again as something that is going to be worth investing in. And that’s called computational thinking. And this is this is an example, maybe one of few where K through 12 really beat us to this. So in in most k through 12 curricula there is a treatment of computational thinking. And they’re working on building computational thinking and students. What it is is essentially kind of what we talked about before, your ability to decompose and structure problems in a way that could be solved by a computer or another human. I often liken it somewhat to expository writing, or even how you would tell a toddler to do something. So if I have a problem, the first step is that I need to think through it, and I need to think about how my toddler is going to process the information and process the problem. And then I need to feed him the problem and instructions in a way that his brain can process so that it can it can get done. Like I would not explain how to do something to him the same way I would explain how to do something to someone else. So it’s very similar. I’m going to explain and give instructions to these tools, knowing how they work and what will help them perform best.

Speaker3: [00:33:30] To give you the result. Yeah.

Dr. Taylor Sullivan: [00:33:31] And so people that are better at that problem decomposition and problem formation and feeding, feeding the AI tools, the right prompts and the right problems and the right way they will excel. And so this is a skill that can be developed. It can be tested for. It’s already being developed in our kids probably today in most schools. But you know when you when you read about it, it’s being it’s on the same level as literacy, math, math, reading, writing.

Speaker3: [00:34:07] It’s a.

Sharon Cline: [00:34:07] Language then.

Speaker3: [00:34:09] Exactly.

Sharon Cline: [00:34:11] You’re taking the human element out of a lot of this, right? The human what? How does taking humans out of. The even the results, I guess. How does that affect? I don’t know. This is a big question. Humanity. We’re talking computers. Thinking, kind of generating.

Dr. Taylor Sullivan: [00:34:37] Mostly predicting the next letter.

Sharon Cline: [00:34:39] Predicting the next letter. For so many aspects of I mean, I don’t even know limitless aspects of our lives. Where does the human aspect become? Not affected. But maybe. Valued. How does it affect a human’s value?

Dr. Taylor Sullivan: [00:35:03] This is a it’s a great question and it’s one that I’ve talked about with colleagues. And it’s also one that comes up a lot when when you go to conferences that are talking about these types of topics. I think the answer is we don’t know. And maybe what we think today may or may not be the same moving forward.

Speaker3: [00:35:25] Truth.

Dr. Taylor Sullivan: [00:35:27] I think that there is still something uniquely human about our ability to empathize. And so kind of that emotional intelligence side of things, social perceptiveness. There are some things that humans will retain value for having, having these abilities. But again, that’s today. It it certainly.

Speaker3: [00:35:50] Could develop as we speak. Yeah.

Sharon Cline: [00:35:52] Well, in terms of the voice over world, having a voice automated this way. Right. So anyone could go into voices.com and if there’s someone’s voice that has been replicated, they can put their own copy in and there are restrictions around it. For example, if I would never do an ad for a gun or a political ad or something, then there are words that I can put in there that will automatically knock out any anybody’s ability to put copy. But there’s just something missing when you listen, I mean, I listen. It’s fine, I think for a medical journal or certain applications that way. But in terms of of taking a breath when you’re speaking, you know, having a catch in your voice.

Speaker3: [00:36:37] Yeah.

Dr. Taylor Sullivan: [00:36:37] The imperfections are what make it perfect.

Speaker3: [00:36:40] Right.

Sharon Cline: [00:36:41] Exactly. Which is so interesting because it’s it’s fast. It’s growing so fast that it’s. I. My brain is like wanting, okay, what I want is for someone to use my voice where I don’t really have to do a whole lot of work, right? So it’d be great. Like you get to pay me for my likeness, so to speak, in the middle of the night, you know? Next thing I know, I’ve got a check. There’s something great about that. But then there’s also an aspect of it where that may not really reflect what I would have normally said, or how I would have really sounded if I had done it on my own, but also. Like I said this, there’s a human aspect of it that feels like you can tell when it’s missing.

Dr. Taylor Sullivan: [00:37:16] Yeah, well, and also, if you read enough output from these tools, you can start to detect kind of patterns and like it just structures responses in a certain way. And like I have a I have a joke with my team. I don’t care if they use ChatGPT to do their work, great. Do it. But I can usually tell if they’ve used it to write something. And again, I don’t care. I’m all for that, but I want to at least know that they know everything that that they supposedly wrote about. And you know, one of our games and our team is to try to find paragraphs that it wrote in, like our white papers and stuff like that.

Speaker3: [00:37:59] So are you able to.

Sharon Cline: [00:38:00] Find them pretty well, though?

Dr. Taylor Sullivan: [00:38:01] I mean, we as humans are getting better at masking its use. And and you know, as we get smarter and we get to be better prompt engineers, we will be able to coach it into our style. We can train it how we would like to.

Speaker3: [00:38:18] Sound more casual.

Sharon Cline: [00:38:19] Or more.

Speaker3: [00:38:20] Formal.

Dr. Taylor Sullivan: [00:38:21] Yeah. And you can, you know, you can have it reply with a certain tone or, you know, emphasizing certain points or give it like a frame of mind. I use it for this a lot for at work if I want to, for example. Understand how a particular buyer persona would react to something. I train it to take that persona, and that might be what are you say you’re a chief technology officer that’s considering buying a hiring assessment platform for your engineers. What are your biggest pain points in in your role? What are you most concerned about? What are your what criteria do you use to make business decisions? What would you be looking for in a tool like this? So I have it, you know, think through its persona. And it’s very good at this. It’s very good at, at at developing personas. And then you can ask it questions. Now that you have your persona, you can say, okay, given all of this, like what’s the best way to sell to you?

Sharon Cline: [00:39:28] And it works.

Dr. Taylor Sullivan: [00:39:30] Yeah. It’s great. And you can also have it. My favorite use for it is just to write emails, and especially if they’re like a hard email that you, you know, you don’t really want to write like I had I had to email my neighbors about having some trees removed because they were at risk of falling on my house. And like, I could have wordsmith that all day. So I just had ChatGPT write it, and it came up with some points that I wasn’t even thinking of. And I’m like, great, I don’t have to spend my afternoon like trying not to alienate my neighbors.

Sharon Cline: [00:39:59] Is it easier to send because it actually isn’t you? Do you know what I mean by that?

Speaker3: [00:40:03] Yeah. Um.

Dr. Taylor Sullivan: [00:40:05] Maybe it sounded a lot better than I would have sounded for that particular.

Speaker3: [00:40:10] No, like, I appreciate that.

Sharon Cline: [00:40:12] But then there’s also, like, some way I’m imagining that conversation. And for me to be willing to have someone else write, it gives me a step. There’s a step between me and the email that I can sort of feel like, well, I mean, I don’t even know exactly what it said. I’m sure it sounded good. You know, like there’s like, no, I’m.

Speaker3: [00:40:29] Not responsible.

Sharon Cline: [00:40:30] For the result of whatever it is you feel. Because I did not write this. Yeah.

Dr. Taylor Sullivan: [00:40:34] You’re on to something there. That certainly is. I mean, I sent it, so I certainly felt confident.

Sharon Cline: [00:40:40] I would have to, though.

Speaker3: [00:40:41] That’s what I’m.

Sharon Cline: [00:40:41] Saying. I’m looking for ways out. Clearly. That’s what I just told you.

Speaker3: [00:40:46] Yeah. No, I mean, be responsible for anything.

Dr. Taylor Sullivan: [00:40:49] Welcome. Welcome to my life. And, you know, all jokes aside, I work for a company that encourages us to use these productivity tools and encourages us to be creative in how we use it, where we use it, when we use it, try using it for everything. Some things that will work well. For some things it won’t. We share prompts with each other like hey, have you tried this prompt before? So there’s a whole community around, like how can we better use these tools? Now? That’s not the reality in every company. And there’s a lot of resistance still. And you know, there’s resistance within individual people, but there’s also resistance in at the company level. Like we’re not going to let our employees use these tools. And there’s good reason for that. You have to know what kind of information you should and shouldn’t put in, like you should never put in your PII. You should never put in secret business information or sensitive data or anything, because all of that technically can go into its training set and be used in the next training, the next model. So just like you wouldn’t give it your Social Security number or, you know, things that you don’t want out of your possession and out of your hands should not be put in there.

Sharon Cline: [00:41:57] We were talking about school and the education system and how they’re a little ahead of us in this way. Us, I say, us the world. Um. All right, so I’m imagining that I am a high school student who did not want to read catcher in the Rye or whatever book and didn’t want to write a paper about it. So I had I write it. What’s the downside of my not having read this book and just been able to get the grade and just move on, like where all of our curriculum at strategic in the way that they’re asking us or asking students to read and learn for critical thinking purposes, I imagine. Yeah, but if you can bypass that, is there going to be a backlash somewhere or downside or blowback when you are not using those critical thinking skills in the way that a traditional student would have?

Dr. Taylor Sullivan: [00:42:47] So this is a hotly debated question. You know, given everything is kind of born in the education system. I joined a Facebook group that is for teachers and ChatGPT for teachers. I just wanted to see how they’re talking about it and like, how are they using it in the classroom and what’s the general like temperature. And so I’ve been in that group for about six months. And even in that six months I have seen a general warming to the technology. So I think they’re not asking the question of what if students use this? They’re assuming students are using it and they’re encouraging students to use it. And so they’re moving away from how do we stop it? To how can we constrain our assignments, how can we help them become better, prompt engineers and more critical consumers of information so that they can detect a hallucination? So I think that’s kind of the direction that we’re going to be going. At the same time, my my sister in law teaches English, and she and I have debates on whether it should be allowed to be used to write essays about books or, you know, express ourselves. And, you know, we have never convinced each other of of the other way. Like, I don’t think it’s cheating to use it. And. It’s an issue like. And I have these debates with colleagues. Now, if if an engineer uses it to complete a coding test, is that cheating or is that like using a calculator? And I heard I was at a conference once talking about AI, and there was a guy in the audience that was a principal. And, you know, this was back in early spring of this year when people were still kind of freaking out about ChatGPT. And this gentleman stood up and said, guys like math teachers didn’t lose their jobs. When the calculator was made. They learned how to teach with the calculator. And just like that, teachers aren’t going to lose their jobs. People aren’t going to. We will adapt and we will find a way to add human value with this tool.

Sharon Cline: [00:45:04] That’s fascinating to me. Also, monetarily, how this affects someone’s job description, I imagine. So if you’re a copywriter, do you get paid as much if you are using your own brain or if you’re using ChatGPT to help you with copywriting? I’m wondering how that affects salaries.

Dr. Taylor Sullivan: [00:45:24] Well, I’ll tell you who is making high salaries are people that are using AI to build things, or people who are building AI systems. The number of job postings for those types of roles has absolutely skyrocketed, and so there’s a high demand for those skills. If anyone’s considering a career change, I guess, I guess that all boils down to is it detectable or do you feel that you need to disclose that you used it? Maybe we’ll start to see clauses like that in in consulting or in contracts where it’s you have to disclose if you used it and how. I would guess that that probably will happen. Now does that mean you should get paid less? Maybe, but I don’t think so. You’re still using the available tools, just as you would go Google some things, maybe to help you write something. This is kind of a supercharged Google.

Speaker4: [00:46:26] Hmm.

Sharon Cline: [00:46:26] What do you see? As? Do you see an ultimate downside?

Dr. Taylor Sullivan: [00:46:35] It’s hard to say because we don’t. This is not the end of this. And I think that there’s promise and peril right now and we don’t know. I think, you know, it depends on your perspective. Like if this matches humans in three years, five years, ten years, is that a downside? Maybe in some ways, maybe not in some ways. I think that. For me and in my world, using it to commit fraud. I think is possibly going to be one of the biggest downsides. But at the same time, it’s almost making us need to examine how we’re defining fraud.

Speaker3: [00:47:27] Interesting.

Dr. Taylor Sullivan: [00:47:28] So is using it to write something? Is that fraudulent, or is it using a tool to help you with something? Or is it is I don’t know, you could go down this path, pull this thread.

Sharon Cline: [00:47:43] But the bottom line, though is that it’s not going anywhere. And there is, like you were saying, there’s a warming that you’ve noticed in teachers because it isn’t. You can resist as much as you want, right? But if it’s going to be here and it is here to stay, then it behooves everyone for. I’m having an attitude of using this for something good, obviously, but also being aware of the way it can be used, because that’s the biggest thing. Like you were saying with fraud. When someone I’ve heard these, these calls, that’ll happen to people where it’s their their mom’s voice, you know, saying, I’ve been kidnaped, you need to send $1 million. Those are the kinds of things you’re talking about regarding fraud, right?

Dr. Taylor Sullivan: [00:48:26] Yeah. And and, you know, this gets into morality and and what makes something ethical and unethical. A lot of that and many definitions of ethics comes down to intent and motive. Me using it to write something for work I don’t have, like, you know, malicious motives there. I’m just trying to make my life easier and get to an end point faster than I would have been able to get on my own. Now, if I’m using it to do do ill or have ill will and befall harm on someone, well, that’s a little bit of a different story. So maybe we view it as a.

Speaker3: [00:49:12] Weapon.

Dr. Taylor Sullivan: [00:49:13] And in the right hands it’s safe. If in the wrong hands it’s not.

Sharon Cline: [00:49:18] But it’s in the hands of everyone right now, right?

Dr. Taylor Sullivan: [00:49:20] Yes. Well, not everyone has adopted it. Not everyone has opened it. I did a quick poll of about 20 to 25 friends, and most are not using it regularly. Some have never used it at all. And you know, going back to this being like the internet, you’re exactly right. It’s not going anywhere. So my recommendation to these people is just try using it for something like log in, create an account, just play with it, see what it can do. And and just kind of have your experience unfold with it a little bit. If you use it and you don’t find it helpful like don’t write it off, try it for something else. Eventually you know, you’ll you’ll stumble upon a use that’s really helpful. And actually that, I think is how I ended up here, is we have a mutual friend, and I was helping her think through how she might tweak her resume and cover letter for certain roles. And I was like, well, duh. Feed the job description to ChatGPT and and then feed it your resume and then tell it to adjust your resume to better align with the job description. And she was like, well, it can do that.

Speaker3: [00:50:34] I’m like, well, I don’t.

Dr. Taylor Sullivan: [00:50:35] Know, try it. And we did. And we sat there and did it together and it did a great job of that. And there’s a lot of people writing about how it can help with job search, and that’s a lot of people are experiencing unemployment right now and are kind of getting a little bit of fatigue being in the market, applying to so many roles, and it can really help take away some of that anguish.

Sharon Cline: [00:51:02] So if someone is listening now who doesn’t know very much about ChatGPT, where would you recommend they start?

Dr. Taylor Sullivan: [00:51:09] So I would go Google OpenAI ChatGPT the way ChatGPT will work. There’s also an app in the App store, ChatGPT you can download it and you can log in with Google or other ways, set up an account, just start playing with it, ask it some questions. I remember one of the earliest questions I asked it because, you know, last time this year we’re approaching Christmas. What should I get my four year old for Christmas? And, you know, ask me some questions. Like what? What do they like? And I was like, music. And they’re like, okay, these types of instruments are appropriate for a four year old. And it was really good. And I’m like.

Speaker3: [00:51:48] Wait, whoa, this is.

Dr. Taylor Sullivan: [00:51:49] Great drums.

Speaker3: [00:51:52] Is that great though? Well, that’s I’m just kidding.

Dr. Taylor Sullivan: [00:51:57] We do have drums and their.

Sharon Cline: [00:52:00] Drums are amazing. I’m a drummer fan.

Speaker3: [00:52:02] Yeah.

Sharon Cline: [00:52:02] Well, that’s so interesting because the way you approach it is just like to have fun. Yeah. Just play. Yeah.

Dr. Taylor Sullivan: [00:52:09] And you don’t have to do it for like high stakes things like have low expectations, have, you know, go in with a curiosity. So and like the meal planning one, I just was thinking like what are the tasks I hate the most? And like if I really wanted to have like the highest impact on my like mental health, what’s the worst thing? So it’s for me it’s meal planning and cooking, but so I’m gonna start there and then it’s every day as I run into a problem. Like if I find myself thinking or like, struggling through a problem for more than five minutes now, I’m like, okay, duh, I’m going to go ask ChatGPT. And, you know, I’m thinking through how I’m going to present it and what I’m going to have it do to best help me. But I don’t struggle too long without giving it a shot.

Sharon Cline: [00:53:02] I’m thinking about how I use YouTube University to help me with lots of things, like something that’s broken and I’ll try to fix it myself, as opposed to hiring someone to do it, which sometimes works and sometimes doesn’t. But I’m imagining I could use this in the same way.

Speaker3: [00:53:20] Yeah, like say.

Dr. Taylor Sullivan: [00:53:21] You have a problem with your computer. You could say, you know, you’re an IT specialist and you need to help a customer troubleshoot an issue. Here are the things that are going on. Provide a list of steps that the user would need to take to troubleshoot the issue. Done. One second later you have some instructions.

Sharon Cline: [00:53:40] Just the speed of everything.

Speaker3: [00:53:42] Too, I know.

Dr. Taylor Sullivan: [00:53:43] And it’s getting faster and faster. It. Well. Caveat. Every now and then the OpenAI website goes down or is just unavailable. And I can’t tell you, I get this like this need this, need to use it. And I’m like, well, how am I going to work today with with this? And then I remember like, hey, it wasn’t so long ago when you didn’t have this and you had to, like, actually use your brain for these things. But but that’s I mean, that in itself is kind of remarkable that it creates like a delay in my productivity when it’s not available.

Sharon Cline: [00:54:22] So recently I upgraded my phone because my other phone was dying and it wouldn’t stay on. It would be on for a second and then would shut down. So I was trying to log in to work and my my VPN wasn’t able to. The new phone hadn’t been switched over, and the stress level that I was under, it took about an hour for me to get everything up to speed, from old phone to new phone, because I kept having to restart. But it really highlighted to me how even just logging into work and having passwords for any of my anything was on my phone, and that the minute my phone wasn’t working right, I actually was like, what do I need to do? I need to fix this today this minute and panicked, kind of panicked.

Speaker3: [00:55:01] Yeah.

Dr. Taylor Sullivan: [00:55:02] I will say I’m not to the point of panicking without it, but I am to the point where I will restructure my day and like, find something else to do because I know I want to use it so badly for whatever the other task was. So I, you know. I will put things off until it’s available and pick up something else that I don’t need it for instead.

Sharon Cline: [00:55:26] My goodness. I have to say I have not used it other than to just play around. I did ask it once, what’s the what’s the best way to become a successful voiceover artist? Since that’s what?

Speaker3: [00:55:39] Is it helpful?

Sharon Cline: [00:55:40] It was. But what’s nice is I actually was doing those things like, you know, school has taught just in the industry, there are certain things that are pretty standard for people to become successful. And so I was really happy to see, well, there’s no out of the box, something that I never would have thought of. I’m like, well, I’m doing everything I can then, so that’s good to know. It was reassuring.

Dr. Taylor Sullivan: [00:55:57] Yeah. Oh, nothing. Nothing is greater when you ask it some things and it tells you some stuff that you already know. I use it a lot to write, like job aids, things that would have taken me forever because I just couldn’t get the words right. Or, you know, I was wordsmithing them to death because I can read it and know whether it’s right or not. But if, for instance, I want to, I don’t know, train someone how to conduct a job interview, like I could write instructions or I could have it write instructions, and then I make sure that they’re right.

Sharon Cline: [00:56:37] It’s a lot.

Speaker3: [00:56:38] Yeah.

Dr. Taylor Sullivan: [00:56:39] Yeah. And I always, I always think like, wow, I have really I really have an appetite for this. And I’ve really embraced it. And like I think it’s so cool. But there is always that burning like, is this, is this right. Like, are there risks here that I just don’t know about? And I’m sure there are. I have chosen to take a really optimistic stance on this technology and really embrace it and be open to it, because I know that our future is it’s not going away and it’s not getting less, it’s getting more. But there is that kind of nagging like little, you know, devil on my shoulder that’s like, are you doing the right thing?

Speaker3: [00:57:21] Well, I think the.

Sharon Cline: [00:57:22] Same thing about social media, because, you know, obviously there’s a whole generation of people who have grown up with it, but I didn’t. And so is there a downside to growing up with social media constantly and being in touch with people constantly? When I was younger, you know, you were home, when you were home, you either had a phone call or you were home. You know, it wasn’t like you were constantly being bombarded with images and thoughts and words from other people and energy from other people. And there’s something kind of peaceful about having a downtime that I chose, I guess.

Dr. Taylor Sullivan: [00:57:55] And we knew our neighbors and we we had we had a community that was actual physical and closer proximity. And I think that’s one of the reasons we’re kind of seeing the death of the, of the concept of a neighborhood because people are finding neighbors online. And I also think that has a lot to do with some of the political polarization we’re thinking we’re seeing, because it used to be that you had to get along with your neighbors, you know, you figure it out, you find some middle ground. But now because.

Sharon Cline: [00:58:26] You help each other.

Dr. Taylor Sullivan: [00:58:26] Yeah, you help each other. But also, like if you and your neighbor have vehemently different political views, you have to figure out a way to be civil and talk about it. Whereas now you can make your own neighbors and they can all have the exact same beliefs as you, and you are in an echo chamber, and you’re never confronted with the need to be civil with people that have really different beliefs from you. So that’s another story for another day.

Sharon Cline: [00:58:49] Yeah it is, but but important.

Speaker3: [00:58:52] Yeah.

Sharon Cline: [00:58:52] An interesting because like you said, it’s moving so fast and we don’t know the impacts of a thousand different things. I don’t know the impact of any of the things that I’m like even fearless formula. So happy to do this but hope it lands where it’s supposed to. But I don’t know the impact long term of anything like that. Yeah, my brain does not work that way.

Dr. Taylor Sullivan: [00:59:09] Well, I feel like that’s a hard way for a brain to work and that’s a big burden. So, you know, I just try to use my best judgment with, with these kind of things. But I’m sure we’re all making mistakes and we’re all doing some things right. And you know, that whole quote about like, sometimes it’s not the destination, it’s the journey.

Speaker3: [00:59:31] Like, none of us know the destination.

Dr. Taylor Sullivan: [00:59:32] And it’s probably really scary and really exciting. Who knows. But, you know, we just.

Speaker3: [00:59:37] Have you.

Sharon Cline: [00:59:39] Have you ever asked ChatGPT what the downside is to ChatGPT?

Dr. Taylor Sullivan: [00:59:43] Yes. Well, I actually I was doing some prep work for, for this and I was asking him about like, what are your limitations? And yes, I do speak to it like a person. I’ll tell it. Thank you. And please, this is kind of like for folks that use it, if I like, share my prompts or my output with someone and they can see they’re like, oh.

Speaker3: [01:00:05] I always say please to chat too.

Dr. Taylor Sullivan: [01:00:07] I’m glad to see that other people.

Speaker3: [01:00:08] Do that.

Dr. Taylor Sullivan: [01:00:10] Anyway, so I don’t.

Speaker3: [01:00:12] Even remember what we’re talking about.

Sharon Cline: [01:00:13] Like, do you ever ask ChatGPT like where where is the downside of having you in my life? Or I.

Dr. Taylor Sullivan: [01:00:18] Do? Well, also, like, you know, in the assessment industry, the first thing you’re worried about is people cheating with it. And I’ve now had this like existential question like, is it ever cheating if you use it like, I don’t know, I mean, if this is always going to be around and if somebody’s going to be able to use this on the job, is using it for this test cheating, I my head goes to these places and I’ve kind of like drawn a semi-artificial line, like if they’re using it to completely misrepresent their skills and abilities and like there is no shred of truth, then. Yes. But at the same time, then we probably need to build better tests that are less susceptible to a non proficient person being able to fake proficiency. Like maybe we just need to try harder and make things less fake able.

Sharon Cline: [01:01:12] I agree with that. I’m thinking how much I use Google for everything. Like I can’t think of this author. You know, they wrote this book that was about this or that. Let me look it up. So is it. Does it mean that I don’t have that knowledge in my head? Does it mean that I’m not smart? It just means I can’t remember at this moment, or what was the name of that character? Or, you know, I get curious about a thousand things. My brain doesn’t stop. So I look up things all the time.

Dr. Taylor Sullivan: [01:01:37] But this goes to what we were saying, though. What you’re good at is knowing what questions to ask, and for knowing that there is an answer and how to instruct something to get it. So when you’re asking Google a question, it’s not like you’re just showing up and staring at Google and waiting for an answer to magically appear. You are still using your brain and your knowledge to know that an answer is there and there is a solution, right? You just need to get to it.

Speaker3: [01:02:06] And it’s not.

Sharon Cline: [01:02:06] Cheating. It’s not a reflection of my intelligence or whichever way you want to phrase that. Um.

Dr. Taylor Sullivan: [01:02:15] I always say like it’s a tool. Yeah, it’s a tool. And, you know, there’s some things that for every, every job, there’s going to be some neat things that you need to know and not have to go to Google for. Like if you were at a networking event and you had to be like, oh yeah, let me answer that question. Hold on, hold on. And, you know, you pulled out your phone to answer like a really basic question about voice overs or something like that. Like not good. But for the most part, we’re not in a lot of those settings where your knowledge needs to be just like right there. I mean, some jobs are like, I certainly would want like a foreign diplomat to have a general knowledge of the world and geopolitical things and not have to pull out their phone to answer a question about some major political thing. So and my husband’s a lobbyist. So we talk about this a lot and he just retains information like it’s crazy. And but I get it. He needs that. He needs to be able to on the spot, you know, roll something off in a in a face to face setting. Um, and so we found that ChatGPT is not going to help in situations where you’re in a meeting and you need to. Refer to something or you need to create on the spot. But you’re right, maybe that will make more face to face interactions though. I feel like the trends kind of in the opposite with remote and distributed work.

Sharon Cline: [01:03:35] Exactly. I was thinking the same thing, like we used to be in offices all the time, and now we’re not more and more. So how how much will we even know? Well, I was thinking a political speech, right? You can’t fake having seen it yourself, right? Someone speaking, it’s not going to be a deep fake.

Dr. Taylor Sullivan: [01:03:53] They probably use ChatGPT did write the speech though.

Sharon Cline: [01:03:57] Right? Right, but they really said it. It’s not someone presenting me, saying something I would never have said.

Speaker3: [01:04:05] Right.

Sharon Cline: [01:04:06] So that will be considered. Um. And the truth, I guess a true experience, a true true words I was there this group.

Dr. Taylor Sullivan: [01:04:14] Like original.

Speaker3: [01:04:15] Content, original.

Sharon Cline: [01:04:16] Original content that you can verify.

Dr. Taylor Sullivan: [01:04:20] Maybe unless deepfakes get really good. But I think the two that’ll be, you know, maybe there’s going to be some space for like, third party verification companies that, you know, their one job is to detect fake from real or to like watermark or certify that things are real.

Sharon Cline: [01:04:38] This is the last question I have for you, because I know we’ve gone over a little bit, but I really appreciate you spending this time with me because, again, I’m trying to catch up and I’m not very articulate this way. So I’m like, I struggle, you know, and I appreciate you being so patient with me as I ask you a bunch of questions.

Speaker3: [01:04:51] Now, like I said, I.

Dr. Taylor Sullivan: [01:04:52] Talk about this stuff all day. And, you know, I again, I’m not an expert in the technology itself, but you.

Speaker3: [01:04:57] Sound like you, I really.

Dr. Taylor Sullivan: [01:04:58] Have. Thank you. Fake it till it matters. Chatty is actually saying all of this right now.

Speaker3: [01:05:04] Yeah.

Sharon Cline: [01:05:04] Really? I will never.

Speaker3: [01:05:06] Tell a deep fake.

Sharon Cline: [01:05:07] I’ll never tell. No. My question for you is, in terms of relationships, how does ChatGPT affect loneliness?

Speaker3: [01:05:21] I don’t.

Dr. Taylor Sullivan: [01:05:23] Well, one quick funny story, and then maybe I’ll try to answer for real. Um, one of the first things that my husband used it for was to write me an email inviting me to go on a date, but it was like this. Absurd. He gave an instructions to frame it in, like, an absurd way. And so it was like, Madame.

Speaker3: [01:05:47] I request the honor of.

Dr. Taylor Sullivan: [01:05:49] Your presence on our back porch tonight to watch an episode of the first 48. It was like this. Really? Like crazy just asking me to go watch TV downstairs. So that was funny. Um, so I think that was a plus. A plus in our relationship was that he used it in a creative way to make me laugh.

Speaker3: [01:06:10] Um.

Sharon Cline: [01:06:12] So a good side, like a but a human side, too, because he used it to make you laugh and like, you obviously knew it wasn’t real or whatever. Yeah.

Speaker3: [01:06:19] That’s cool.

Dr. Taylor Sullivan: [01:06:19] Yeah. Um.

Speaker3: [01:06:22] I think in some.

Dr. Taylor Sullivan: [01:06:23] Ways it could help with addressing loneliness. I mean, it makes me think of that movie her where Scarlett Johansson. Yeah, exactly. I mean, I think that this will take some humanistic traits and make them accessible to people, so it may make them feel like they are less lonely. But at the end of the day, it’s a large language model. It’s not a human. Maybe having so much access to information at your fingertips might help with connectedness. Just being able to learn something quickly or get certain kinds of information quickly may help people feel less isolated. If if they don’t know something, they can find out. And so maybe just having access to information could help with that. But yeah, I mean, you know, that emotional intelligence, social connection, empathy, actual communication, non-verbals that’s obviously not going to be there today. Maybe it will be in the future.

Sharon Cline: [01:07:32] We just don’t know.

Speaker3: [01:07:33] We don’t.

Sharon Cline: [01:07:34] We’re living it as we speak.

Speaker3: [01:07:36] Yes.

Dr. Taylor Sullivan: [01:07:37] It’s I mean, it’s an exciting time. It’s an exciting time to be alive. It’s been really fun to watch this roll out. So obviously my eyes every day are focused on the software engineering industry, and it’s kind of hitting it in waves, but it’s changing everything. So it’s just some companies are changing faster than others.

Sharon Cline: [01:08:05] So, Dr. Taylor Sullivan, how could someone get in touch with you? If they would like to know more about what we talked about or anything that you’re doing in your business world?

Speaker3: [01:08:14] Yeah.

Dr. Taylor Sullivan: [01:08:15] Well, you can email me at Taylor Sullivan at codility. Codility.com.

Sharon Cline: [01:08:22] Excellent. Thank you.

Speaker3: [01:08:24] Yeah. It’s my.

Sharon Cline: [01:08:25] Pleasure. This has been very, very interesting for me. And I was actually going to see if maybe in six months, if you would be willing to come back and we can speak about this again and just see what has changed.

Speaker3: [01:08:37] Yeah.

Dr. Taylor Sullivan: [01:08:37] I’d be happy to. I’ll send my deepfake next time.

Sharon Cline: [01:08:41] I’ll send mine.

Dr. Taylor Sullivan: [01:08:42] Perfect. We can go get it. We can go get a glass of wine instead.

Sharon Cline: [01:08:49] Sorry. That’s funny. And kind of sad, but true.

Speaker3: [01:08:52] Who knows what will be.

Dr. Taylor Sullivan: [01:08:53] Or that’s the most exciting thing ever. Like, I would love to imagine a day when I can just sit on my couch and, like, just chill and my house is doing everything for me. My job is covered. My kids are somehow cared for and entertained.

Speaker3: [01:09:07] Cancer has been cured.

Dr. Taylor Sullivan: [01:09:09] Yeah, yeah, yeah. Maybe that’s where we’re going. I like to tell myself that.

Speaker3: [01:09:13] I believe.

Sharon Cline: [01:09:14] That too. And I like that you talk about a positive spin here, because in my mind, it’s. It’s like this world that seems scary. Not. I don’t know if scary is the right word. Unfamiliar. I don’t even know what this is going to do. I always think dark, like, oh no, but you’re talking about something as being really positive and helpful and encouraging you to make different connections in different ways, whether it’s in your business or interpersonally with your children. I love the idea of that, where you’re using it for something good. Super power for good.

Speaker3: [01:09:45] Yeah.

Dr. Taylor Sullivan: [01:09:46] Well, I mean, psychology tells us that unfamiliar things are scary usually, and the best way to deal with that is to confront it and become familiar so that, you know, my advice stands download it. Just try it.

Sharon Cline: [01:10:00] See what happens. Yeah. Make some dinner with five ingredients.

Speaker3: [01:10:03] Yeah, I probably should.

Dr. Taylor Sullivan: [01:10:05] Actually make.

Speaker3: [01:10:06] Something.

Sharon Cline: [01:10:06] I’m going to try that actually I’m going to I’ll text you and let you know how it goes. Yeah.

Dr. Taylor Sullivan: [01:10:10] Great. No that’s one of my favorite things. Anytime I like teach someone how to use it, I always am like, send me your output.

Speaker3: [01:10:16] It’s like.

Dr. Taylor Sullivan: [01:10:17] God.

Speaker3: [01:10:18] I’m such a dork.

Sharon Cline: [01:10:20] No, you’re amazing. And I really appreciate the insight. I never would have been able to understand everything that you spoke to me about by just reading it, or even having ChatGPT tell me. So. Thank you for your presence today in the studio and helping me to understand what this new frontier.

Speaker3: [01:10:35] Yeah. My pleasure.

Sharon Cline: [01:10:36] And come back.

Speaker3: [01:10:36] Yeah, I’ll come back.

Sharon Cline: [01:10:38] Okay. Thank you. Thank you all for listening to Fearless Formula on Business RadioX. And this is Sharon Cline again, reminding you that with knowledge and understanding, we can all have our own fearless formula. Have a great day.