Generative AI: Productivity & The Fight for Reality

Speaker A: Foreign. Today on Fintech Bites, we're unpacking one of the most influential business technology topics of the decade, generative AI. Inspired by McKinsey's exploration of what every CEO should know about gen AI, which be digging into how this technology is reshaping business value, risk and strategy, especially in fintech. We'll ask the big questions leaders are wrestling with and explore what students and early career professionals should understand if they want to thrive in the AI era. What makes generative AI fundamentally different from other types of AI and why does that matter for businesses today?

Speaker B: Well basically Genai catapulted artificial intelligence from academia to, to mainstream adoption within three years, actually within five days because we went from no users to 1 million users in five days, which is a record by any standard because I, uh, Instagram took three and a half years to do the same thing to get to 1 million users. So why did it happen? It happened because AI or Genai or which is most commonly associated with AI, is that it made it easy to just ask a question and get an answer. And that's what people want. People don't want to spend time going to Google Search, typing in a uh, query, looking at links and trying to read through that and go to the next link and the next link, ChatGPT just give you the answer instantly. So the reason that it became that popular is because it made things easy and the user experience in all of this is excellent. It gives you what you need when you need it without going through too much. And I think it should be adopted across businesses and everywhere. And it will happen with or without. People are going to use it just like they use it before when uh, they were using Google Search. People want convenience, they want instant answers. They want to know what they're looking for and get an answer right then and there. Now of course, San GPT has the issue of hallucinations, which makes it difficult to sometimes be misled about certain things that sound real but are, uh, not. But it's only a matter of time before this gets improved. But in the meantime it's a great tool to use in any aspect of business.

Speaker A: I mean Genai isn't just automation. It creates new content, text code, images rather than only analyzing data. So it operates on foundation models like Chi GPT4 that can be tuned to diverse tasks across workflows. This broad capability accelerates transformation but also introduces new risks. Can you just explain for our listeners what uh, ChatGPT stands for? Because I think a lot of people might not know.

Speaker B: Well, ChatGPT stands for the transformer aspect, where it learns the large language model. So the transformer kind of sounds like the transformer movie.

Speaker A: Hmm.

Speaker B: But it's not. So basically, large language models get trained on data. And basically what it does is it asks a question, which is usually the person asking the question, the prompt. And then chatgpt tries to figure out what's the next logical word in that sentence. Now, the LLMs that are available now are based on text. So we're looking at the next generation of LLMs, which is going to be spatial 3D more, where you can see perspective. Because one thing in artificial intelligence, it has difficulty discerning perspective distance. So spatial large language models are gonna address that. But it's still at the beginning stages. So the thing about ChatGPT is at the beginning it's just gonna give you text and try to figure out what the next word is. And it could be good or not so good, but it's not just one word. It can be a whole project, it can be a whole article, it can be a whole summary, whatever it is. But the thing is that we still have the hallucination problem, which is quite a setback when you want to rely on the information that you're getting. But it's getting better and it's only a matter of time.

Speaker A: Yeah, I think, um, Dall E, the other image generator, is inspired by Wall E, the movie. The name. I read that recently.

Speaker B: Well, it's, um, true Dali, as we speak about. It's only three years ago and it's like old news because now we have Narobana and we have ChatGPT, you know, 5.2. And um, and then you have other platforms that, ah, are competing. Uh, it sounds crazy that it was only recent, but we were making very big progress, very big strides. And I, I can tell you one thing. I went to one event where they were talking about, and this is what, six months ago, they were talking about generative AI. They were talking about AI to movie, to, you know, text to video and all that. And there was at this conference in Brighton, uh, they talked about, uh, how it changed everything. People are in special effects and computer generated imaging and so on. The ones that create the avatars or, uh, movies and so on, they're faced with a dilemma because their jobs is at risk. Somebody can do it for them. And I can tell you one thing. When I saw what can be done by a, uh, speech prompt, so you don't even have to type it in, you just say, make me a video. You put a Picture up. You say, take this picture, black and white, picture, whatever. Make it a movie. And it has to be scary, it has to be dark, it has to be mystical, it has to be this, that, whatever. And it did. Now the funny part is some of it was excellent. You had like, what, what on earth did just happen? I mean, within seconds you're looking at a movie made from a picture, but then you saw mo the movie and some of it didn't have the perspective, so it looked wrong. Like, for example, a drawer opening up and, you know, the, the door just came off. There was no drawer. Uh, you know, little things like, you know, too many fingers holding up the drawers. It's six or seven fingers instead of five. Like crazy stuff, you know. So this, this little fine tuning that needs to happen. And I'm sure it's already there. Um, I'm sure, like if you work in Hollywood, you probably have the latest, you know, AI generated content, video, whatever, and then it gets better and better and better. But at a certain point you will just talk into an app or platform and say, make me, you know, the next successful Disney movie. And it has to be very entertaining, very successful and interesting and it will do it all for me from scratch. So the people that are working in that industry, making a living will most likely unfortunately lose their jobs and the AI will create everything for you. And since we are used, accustomed to look at animated movies and we love it like uh, you know, Snow White, all those movies, uh, Avatar, whatever, we won't object to seeing a movie made by AI but just like anybody, we have our favorite actors, actresses, and so we want to see movies that are made by actual people. So again, movies will be made with actual people, but the background is not going to be realistic, it's going to be AI generated. So, uh, if you are a producer in Hollywood making a movie and you need to film in the Swiss Alps or anywhere else, instead of going there physically and spending fortunes bringing the team out there, the film crew, the uh, the tech crew, everybody there. You can just do all of that in your studio in Los Angeles and you can actually record the whole movie, the whole scene, staying in the studio in Los Angeles or Hollywood, whatever, so you don't have to go anywhere. So it will make movies cheaper. But I think people want to see the real background. They want to see really uncharted territory. That's why they have in the movie industry, movie scouts that, uh, find places that people don't usually go to. Very unique, very special. And you want to see that and that's not going to happen. Because the worst problem is you're going to ask yourself, when you see a movie, is this real or not? If you see something, you'd be like, is this the real background? Is this a real city? Is this a real town? Is this a real village? And you won't know the answer unless you read the experts that have been informed about it. So it's tricky.

Speaker A: Yes, I think now it's just you start to dot everything you see and you're second guessing if it's real or not.

Speaker B: The second guessing part is the worst because every time you see something and you like it, you're second guessing and you're like, is this real? Is this being recorded? Is this AI and all that. And I think the biggest issue that we need to talk about is that in the future, and I, uh, find it horrible actually to tell you the truth, is like, whatever you see, you will ask yourself, is this real? And okay, if you see a movie, you know, you're looking at a movie, but if you're looking at the news and something is happening in the world, you'd be like, is this real? Is this actual footage? Because now you have fact checkers by the big media companies, BBC, uh, Sky News, cnn, whatever. And they check the facts, you know, and they even came across fake AI and all that, and deep fakes and just, you know, propaganda, manipulation, whatever. So if, if you have to second guess everything, I, uh, find that horrible. So we're going to come back to the Roots, where you're going to want to deal with people you can actually see up front. I think that in the future a lot of people will not necessarily go to movies anymore, but want to go to a theater and actually see people on stage, not even robots. And concerts are going to increase in value. And I think you can already notice it by the prices going sky high and unaffordable for most people because people want actually to see somebody on stage, dancing, performing, singing, whatever. So a lot of people are going to go back to the Roots. And I think we should do that because we're humans and we want to see reality, we want to see emotion, we want to see things. Hey, we even want to see live bloopers. Bloopers, Excuse me, Bloopers. Where, you know, you see a, uh, concert performance in the lead single falls off stage. Now it sounds bad, but, you know, it's real, you know, and that might bring back home the fact that people are not perfect, make mistakes, might trip, might fall, might say something wrong, might forget or text, whatever. And I think that's going to add to the excitement of seeing something be concerts, performances, musical theater, football, sports, F1, whatever. People want to see genuine human beings doing something in real life.

Speaker A: How should CEOs think about the value of generative AI as a productivity tool, a strategic differentiator, or something else?

Speaker B: I would say all three. Number one, the productivity is baked in. So you actually would be not smart of not using AI to increase your productivity, which by now is free of charge. You could use ChatGPT to do it, but if you want to get a little bit more professional, you could, you know, increase it dramatically. You also need to look at where it's going to be in a couple years time, because now we're still at the beginning stages and I think in not even a year's time, it will get so much better. So you need, as a CEO, to explain, first of all to everybody how you want to incorporate it into your business and alleviate the fears of those people that you're telling them that they will lose their jobs. Because that's what people are asking is like, okay, I start using AI, explain to AI what I'm doing, how I'm doing this and that, and one day the CEO will come, uh, out. Thank you for informing my AI bot here how to do things so you can move on. We don't need you any longer, and that fear needs to be addressed.