Power Lunch - Power Lunch Special: The Power of AI 5/26/23

Episode Date: May 26, 2023

Artificial Intelligence is what everyone is talking about right now – from boardrooms to living rooms across the country and world.What everyone wants to know: is Microsoft or Google winning the AI race? Is ChatGPT going to replace my job? And does AI pose a real threat to humanity?We’ll discuss all of those topics and much, much more – including the industries already being disrupted, potentially changing life as we know it. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Good afternoon and welcome to Power Lunch, everybody. Alongside Kelly Evans, I'm Tyler Matheson, and today we have a very special hour of Power Lunch Plan for you, focused on artificial intelligence and the impact it is having on business, the markets, and society. AI is what everyone is talking about. In boardrooms and in living rooms, everyone wants to know, is Microsoft or Google winning the race? Does AI pose a real threat to humanity? Is ChatGPT going to replace my job? Here's what some of the biggest names in business have said.

We're moving from the autopilot era of AI to co-pilot era of AI. So if anything, I feel, yes, it's moving fast, but moving fast in the right direction. These are new tools, but it's important to know that the existing rules still apply. There is no exemption or legal shield that AI enjoys. We believe that we can and must work together to identify and manage the potential downsides so that we can all enjoy the tremendous upsides. I am personally skeptical of some of the hype that has gone into artificial intelligence. I think old-fashioned intelligence works pretty well.

It can do all kinds of things. And when something can do all kinds of things, I get a little bit worried. And because I know we won't be able to uninvent it. This is a business story. And when you think about AI, empowering people to do business. AI is the next wave of computers. That's why I called it the iPhone moment. We've taken IT and we put it into the hands of literally every single consumer in the world.

This is such a big deal. The introduction of large language models, artificial intelligence, is going to create a productivity boom that we've only seen a few times in the last 75 years. The widespread adoption of some of these tools could lead to Drax. drastic changes in the demand for workers, to wages, to the income distribution. I don't think you can recreate a Larry David or a Seth McFarlal. I'm saying my client or Jim Brooks or Adam, I don't think that's that easy. Well, the advent of artificial general intelligence is called the singularity because it is so hard to predict what will happen after that. I think it's very much a double-edged sword.

I think there's a strong probability that it will make life much better and that we'll have an age of abundance. And there's some chance that it goes wrong and destroys humanity. We will cover all of those issues and more. But first, let's try to tackle how you can best make money off of this AI boom. You've heard about Microsoft, Google, and lately, NVIDIA, those companies at the forefront of the AI push. And the stocks have been rewarded for it. look at those year-to-day gains. The combined market cap of those three names is right around $5 trillion. But there are lots of other opportunities to play this space. Let's bring in Gil Luria,

senior software analyst with DA Davidson. Also, Chris Rowland, senior analyst with Susquehanna, and our very own Steve Kovac. Welcome to all of you. Gil, you are one of the first analysts to say that Microsoft's investment in AI is a big deal. And what would you say to investors now? Have they missed the move? Not in Microsoft. Microsoft has been preparing for this for a long time. As you put this in the context of other big technology wave, let's not forget Microsoft initially missed the move to mobile, initially missed the move to cloud.

That's why Mr. Nadella took over in 2014, and he has transformed Microsoft specifically for this moment in time where that next big wave in technology comes in. He has transformed Microsoft in order to capitalize on it, and they are just getting started. So even Microsoft, Gil, you think, is a name people could still look to position in. What are the others? Are they all the obvious candidates?

Or are there some that you think have better potential to kind of catch up here? Well, I think the obvious ones have played out. That's Microsoft and Nvidia, the certain extent C3 and Palantir. On the other side, Chegg, which is those are 2,023 events. Those companies' benefits or detractions are going to come. this year. For everybody else, the race is still starting. And again, it doesn't always matter who has the best position for the race. It matters how they execute. But as we look at some of the smaller midcaps that are interesting that could capitalize on this, we look at companies like

Zoom Info, like JFrog. Those are some companies that are smaller that already utilize artificial intelligence and are positively leveraged as more code gets written, as more sales calls get made to leverage this technology. Chris, we know about Invidia. We know about a couple of other companies that I guess I could put it this way are building the plumbing, the back office of AI. Give me some other names that are sort of beneath the surface and are going to be part of the structural architecture of artificial intelligence as opposed to being sort of the front end that we interact with. Yeah, since that historic, truly historic report from InVidia this week. Investors are now playing catch-up. They're trying to

figure out the second-order effects, these second-tier companies to benefit. And we are suggesting networking to connect all of these GPUs and AI cards together, takes plumbing, takes networking, and optical companies and networking companies. For example, broad companies. is an excellent beneficiary, we believe, of connecting all of those AI processors together. Chris, it's Steve. I got a question for you about some of the other big tech names that we're leaving out of this conversation. I notice no one on this panel is mentioning Apple, no one's mentioning Amazon. Are they just too late to this? They're a little late to this. They haven't emphasized it in the same way that others have. Google has been really a leader in all of this, and they understand the existential threat

for search for something like chat GPT, for example. It's less clear how someone like Amazon is going to benefit from AI in the same way. So I think they're making appropriate investments, maybe Apple's perhaps a little bit behind the curve here, as they could be ratcheting up their AI capabilities. But generally, each company, I think, is making a massive investment, which we've really seen, as of this last quarter of NVIDIA, suggests that when ChatGTPT was launched in November, all of these orders from all of these other companies came pouring in for these chips. And now they're just starting to show an NVIDIA's guide. Gil, can I come back to you with specific reference to Microsoft? and my observation is this.

The history of legacy companies, and I'll consider Microsoft the legacy company in the software world and in technology, the history of legacy companies going on to dominate the second and third or fifth chapter of whatever the evolution is, is relatively rare. It wasn't NBC or ABC or even Disney

that dominated streaming. It was Netflix, a new company that did that. So why, explain a little more to me why you think Microsoft not only won't fumble this time, but will lead this time? So I'd actually put Google in the position of the incumbent here, just like Xerox invented graphic user interface. And then Microsoft and Apple took over the operating system market. The artificial intelligence made the most progress in Google's back. in the labs of Google, and yet they're slowest to commercialize it among the big companies,

specifically because of the intervasive dilemma, because their search business is so good that they have a huge incentive not to disrupt it. So they're the incumbent. Microsoft, again, has carefully positioned itself to be the disruptor in AI, and they are so far ahead of Google in monetizing it and commercializing it, that they have a big head start. Again, that doesn't mean they'll ultimately be the only winner, But they have a big head start because they've been thinking about this for a while.

Their GitHub co-pilot was launched in 2021. Well before Chad GPT, they've been thinking about this for a while. They've been planning for this moment. And right now they're executing very well. All right, Gil, thank you very much for your perspective today. Chris Rowland, thank you. Steve Kovac, you're going to stick around. I'm sticking around.

We can't get rid of Kovac. Why would you want you? We can't. All right, let's switch gears and broaden the discussion out to AI's impact on the economy and the market more broadly. To do that, we're joined by Peter Bereson, chief global strategist at BCA research. Peter, welcome. Good to have you here. We've talked just a moment ago about some of the immediate corporate beneficiaries of artificial intelligence,

whether it's in those plumbing companies that I referred to or the Googles, the Microsofts, the Amazon, and so forth. What are the other industries or companies derivatively that you think will benefit most? have the most to gain from the AI, from crossing the AI frontier? Well, it's interesting. What we saw in the internet boom was that it took a while for companies to figure out how to monetize the internet. In fact, the productivity gains from the internet generally came before companies really were able to make a lot of money.

A big productivity boom started in the mid-1990s and ended around 2005, just to when tech profits were really starting to rise. So we could see a similar story this time around that it takes a while for companies to figure out how to monetize AI. And that wouldn't surprise me at all, because if you think about where the profits are coming from right now, it's coming from companies that have huge moats around their businesses, companies that benefit from network effects. People use Instagram, because others use Instagram. I don't use Chad GPT because Others use chat GPT. I use chat GPT because it's good.

If something else were to come along, that's better. I would use that. And so it's not at all clear whether the moats around the AI business are going to be so why the companies that currently are leaders in the tech sector will be able to continue to generate the sort of profits that they've been enjoying over the last decade or so. What would you add to that, Steve? I guess my other question for Peter is there was this debate that kind of was going on in Europe this week as they're talking about their own legislation. And it's this idea that part of

the regulation is they need to be more transparent about the models they're training on while we hear, and we heard some resistance from Sam Altman, the Open AI CEO, to that, and they don't want that kind of regulation. I guess my question is, can an open model work, like, such as what Facebook is working on or meta is working on? Can that supersede this? There's been some internal discussions I know within Google that these open models are actually more of a threat than whatever Microsoft is doing. How do you think about that? Yeah, I think that sort of speaks to the point that if you have a moat around your business and you move to an open source model becomes a lot more difficult to preserve that mode. And ironically, the other issue, which I think

needs to be addressed, is that a lot of tech companies today make money because their offerings are great, but not perfect. So if you think about Google, for example, let's say I want to go to Houston, and so I put in a search, you know, how do we get to Houston? Google is going to give me a bunch of links to various airline companies and so forth.

If Google had a chatbot that would tell me, here's the cheapest flight that you would take, take it on this date, well, that's great. That's even more helpful to me. But how does Google make money off that? If Google's chatbot is ethical, it's going to give me the correct result.

They'll tell me which airline is the cheapest. The airline that provides that service doesn't have any incentive to advertise with Google because Google is going to choose them anyway. And those airlines with higher fares aren't going to get any coverage at all. So again, how do you monetize this? It may be actually much more difficult than people realize.

We have to go, Peter, but have you changed your odds from 50-50 chance that humanity gets wiped out by AI since the last time we spoke? No. In fact, when we spoke shortly thereafter, Jeffrey Hinton, who, of course, worked at Google, one of the great pioneers of AI, won the Turing Award, so the Nobel Prize for Computer Science. He said very much the same thing.

So my odds are still high, and maybe because my webcam is really blurry, AI is trying to prevent me from spreading this message. If it can fix your webcam first, then I'll say, okay, 50-50 odds. it wipes out humanity. Peter, thank you so much. We appreciate it. It was great. Don't even worry about it. Peter Bereson, joining us from BCA Research. There's much more to come on this Power Lunch special, the Power of AI, including the search for the next big thing. How did venture capitalists know which companies are the real deal and which are just science

fiction? Plus, the industry's already being disrupted by AI. And how that will change everyday life as we know it, from Hollywood to the menu at your local restaurant. This Power Lunch special continues after the break. Welcome back. As AI evolves into the buzzword of the year for tech, it also brings the risk of AI washing or the practice of throwing the word onto any product or company trying to get funding and attention. Here's what Databricks CEO Ali Gozzi told Dear Jabosa about the questions investors need to ask. Just because some company is doing generative AI and they're slapping on a large language model in their name and they're doing something, but it doesn't actually mean that that's a moat. Why can't anyone else also do the same thing?

So a little of the critical questions that I think you have to ask yourself. It's sort of like just before 2000 IQ bubble. So how do VCs know when the tech is real? Let's ask one. Duncan Davidson is a partner at Bullpen Capital. He's been through a few hype cycles, Duncan. Welcome. What's the first question you ask in trying to separate the reel

from the science fiction? We don't answer that question. That question's a trap. You heard the person just say that it's easy to wrap around chat GPT and throw a business out. Well, if it's that easy, it's going to be 20 more doing it before you know it. And so you don't want to go into that game. That's just fratricide. Instead, we ask a different question. Are you building a real business? And I can tell you more about how we look into that, if you'd like. Yeah, go ahead. Well, first of all, there's a situation that always happens called a platform shift. We went from client server to cloud. The client server winners got smoked by the cloud people like Salesforce.com.

So right now, one question we're asking is, will an AI-based CRM smoke Salesforce? It may seem improbable. They're so big and strong, but this happens every time the platform shifts. So that is number one. Number two, we look underneath it to see whether it's enterprise ready. You know, a lot of these little wrappers around chat GPT that are getting spun up right now are no way ready to go under an enterprise. They're just too lightweight.

So are you trying to replace an existing workflow in enterprise and you have a platform that can go right in there? Now, if you understand what I'm saying, we don't look at the core chat GPT and all that. That's not the business we want to invest in. That takes so much capital now. It's out of the VC range. The analogy I'd give you is when Amazon and Azure built these giant cloud platforms, the VC stopped investing in the cloud platform.

It wasn't our game anymore. It was investing in things on top of it. And that's what we're doing right now, things on top of the AI platforms. So let me make sure I'm understanding you. You would not be investing in the same kinds of things that Google is investing in, that Microsoft is investing in. That's out of your scope. and so you'll invest in things that are more bite-sized or more digestible or where you can have a greater influence.

Or is the analogy sort of the Uber to the iPhone? Like the iPhone was a platform that gave rise to Uber as a possibility? I think you're both right. We're not going to try to compete with Google and Microsoft, whatever, OpenAI. But we are going to try to use that as a platform to compete on the layer higher. Like, you know, obviously when iPhone came out and got to the right format doing apps for a while, was a hugely interesting game to play in. So we're now in that part of the game.

We're trying to invest above the cloud platforms and the AI platforms and above the really expensive tech. And so then the obvious question without asking you to open your playbook 100% is what kinds of companies does this process lead you to right now? I think three kinds. Number one is the platform shift where you're going after a real existing category enterprise software with AI power underneath it. And you go into existing workflow of enterprises. This is the CRM example, but there's dozens of other enterprise software categories where AI can

really empower you. The second one is a data category. So let me give you a problem with these big platforms. You don't know about hallucination. AI can hallucinate, meaning you can ask Chad, GPT, a question, have a conversation and it starts making stuff up. That's just not true. Okay, let's say you want to invest in a medical AI system. You don't want that bot to make things up. So to get around that, you have to have the domain they look at, so-called operating design domain of data, be very precisely defined, and you have to manage the technology to work on AI in that domain. So it's highly accurate because too much is at risk if it makes something up. Those type of database deals where they have a data category that you really understand and own,

those are really good to invest in too. Well, maybe say, go ahead. Quickly, Duncan. Your third point was what? I'm sorry, we interrupted. No, I'm just saying that the third point you can invest in out to the platform shift and the data thing is you could look at this game of wrappers and people doing API calls and AI. And you often find a category which is not going to. going through this huge fratricide, everybody doing it, that is you have a unique knowledge of a

certain category in the market. Those would also look into where the company has a business. So you sort of ask the question this way. Instead of asking, is it a good tech, you would ask, when the hype goes away, is this a good business? And those are some of the categories we'd be looking at right now. Smart way to think about it. Duncan, thank you very much. We appreciate it. We'll have you back soon. Very good. Take care. You bet. Have a great weekend. All right. Still ahead. You want fries with that? Artificial intelligence now improving accuracy for your drive-through order. But can it throw in some extra ketchup packets, too?

We'll find out when this power of AI power lunch special continues. Welcome back. Stocks are higher across the board right now. The Dow in the Green for the first time this week by about one full percent or 300 points. The NASDAQ continuing to soar up 2 percent today on top of yesterday's big Nvidia rally. And today's big winner is Marvell. technology. Technologies. The company says demand for generative AI, boosting its products. It expects AI revenue of about $400 million this year and double that. Next year, check out some of the AI-related ETFs. Only a few of them are out there, but according to our partners

at Track Insight, they are seeing steady inflows and no surprise, big gains for the year. Still ahead from mind-bending special effects to mimicking a singer's voice note for note. AI has already entered uncharted and in some cases unwanted territory in the entertainment business. We'll ask industry insiders how they're dealing with it when this power lunch special, The Power of AI, returns. Welcome back to a special Power Lunch, the Power of AI. We're looking at some of the industries being disrupted already by AI, and one example is entertainment. But for all the things AI can do, here's Academy Award-winning producer Brian Grazer at the CNBC CEO summit explaining one thing AI can't do. Look, the movie and the TV series Friday Night Lights, that came about because I got cut from high school football in front of 250 kids.

And that was a disaster. That was seismic to me. Now, I don't think that can happen to a machine. They're not going to get cut from high school football. And they're not going to feel humiliation, and they're not going to look around the room and look at all those faces and eyeballs that are looking at you and penetrating with penetrating eyes. That's just not going to happen.

So they can't, you know, AI can't have the experiences that brings life to bigger and other ideas. That was beautifully expressed. For a look at some of the ways that AI is still disrupting Hollywood, let's bring in Julia Borsden. Julia? Well, Kelly, it's been nearly a month since the writer's strikes started, and use of AI is one of the key sticking points in negotiations. The Writers Guild is demanding that the studios agreed to, quote, regulate use of artificial intelligence on MBA-covered projects.

Those are projects covered by the terms of this deal. They're saying AI can't write or rewrite literary material. It can't be used as source material and that the covered material cannot be used to train AI. writers raising concerns that AI is a plagiarism machine trained on published or protected material, and writers won't get the credit and payment they deserve. Now, as for the studios, AI could become a huge money-saving force, used to streamline everything from casting to location scouting to special effects and even marketing of films. But actors ahead of their Guild contract expiring are concerned, the Screen Actors Guild telling us, quote, informed consent and fair compensation in the use of AI is a fundamental baseline for the union,

saying performers must be fully informed of all of the details of the intended use for any consent to be valid. And there must be recognition that for performers, their voice and likeness are their stock and trade, and must be fairly compensated for any use. Now, I demoed this technology from a company called Deep Voodoo that enabled someone else to effectively put on and weigh. my face. There's another technology from a startup called Wonder Dynamics, which very quickly turned me into a robot and some other different characters. Now, both of those two startups are working not just with studios, but also with actors to make better-looking movies at lower costs. You can see why the potential to use actors' likenesses, if they weren't participating

and in agreement, could raise some major red flags. Back over to you. Julia, thanks very much. And let's talk about that and more in the specific context. of the music industry, which faces some of the same issues as AI has been used to create believable sound-alikes of some popular artists. And now music labels are saying AI music generators are a direct threat to the industry. For more on this, let's bring in Harvey Mason, CEO of the Recording Academy and a veteran of music business attorney, Peter Paterno, a partner at King Holmes, Paterno and Soriano. Harvey, let me begin with you. What is the risk and what would the potential damage be to the creative community?

If, for example, I were able to interact with an AI generating music machine and say, I would like you to write me a song in the genre of Tina Turner. I want it orchestrated the way Tina's songs were orchestrated, and I want it to sound like Tina Turner. Well, first off, that's happening and it's going to happen. And of course, there are risks associated with it. We need to figure out how the original creator, or in that case, Tina, would be compensated. But it's going to happen.

AI is here. People are using it. Creators are going to deploy it to be more creative. Ever since the evolution of darn near the drone, technology has changed the way people create music. And I don't see it being any different with AI. There are some nuances that hopefully we're going to discuss and need to be talked about. but it's a new version of technology that's going to enable creators hopefully to do more. I'm seeing everything from people being terrified to death about AI to people being extremely excited about the potential of AI.

And it's all over the map. And we can talk about the reason. Could it hollow out the creators, the human creators? I mean, there is, as Brian Grazer said, there is no way that a bot can have the human experience that creates the kind of plaintive sound, emotional valent that real human-created music has? Well, personally, as a creator, I believe that's absolutely true. And at the academy, we are working to find ways to make sure that remains the case. But emotion, heart, passion, feeling, storytelling is always going to have a human component to it regardless of what technology does. Peter, is this back to the future

for you, having worked, correct me if I'm wrong with Metallica, I'm going after Napalph? back in the day? I mean, is this the time to kind of protect copyright really forcefully? Well, I mean, you should always protect copyright forcefully. I mean, we have to reward the creators, people that make this music that everybody listens to and enjoys. So, yeah, the copyright issues have to be sorted. When we say they have to be sorted, so let's take the example of the Drake song, or pick one, pick any one of your clients. If someone else created music using their likeness, would you shut it down? Yeah, I mean, that's, that's, there's two, there's two things going on here.

One is the use of the AI to just create music that sounds like somebody. I don't think there's anything legally that stops that. But if you try to pawn it off as, as somebody else's Tina Turner or something like that, you can't, you can't do that. So, so Harvey, let me come back to you and pick up on what Peter just said. In the case where somebody says, go back to the Tina Turner, Turner example, I want to create something that sounds like Tina Turner. The estate of now of Tina Turner, rest in peace, could bring action against that creator of the

artificial one. But the artificial creator can always say, but it's not Tina Turner. It just sounds like Tina Turner. It's just orchestrated like Tina Turner. But it isn't the same, just the same as Joe's Cola isn't Coca-Cola. Under the current conditions, there is no clarity for it. We have to establish a way for the original artist, if their name and likeness is going to be used, for them to be remunerated properly.

Right now, it's a little bit of the Wild Wild West. People are going to be doing that. I would think in the future, Tina's estate, probably not right away, but at some point is going to be excited to be able to have Tina's voice used in new ways as long as people are crediting Tina the right way and people as consumers on understand there's a differentiation between an AI version of Tina and the original team. Interesting. I agree.

At some point, it could actually create, you know, streams of revenue for way into the future. Absolutely. Huge revenue. Huge revenue and huge, more information for consumers and fans of music. Kind of the switch from Napster to, you know, Spotify or whatever. Thank you both, gentlemen. We'll leave it there for now. Harvey Mason and Peter Paterno.

Coming up, a company using AI to solve one of the biggest problems facing humankind. getting the wrong food at the drive-thru. We'll show you how it works and talk to the company's CEO. That's next on this Power Lunch special, The Power of AI. Welcome back, you know, it's happened to everyone. You pick up food from the drive-thru, you get home or you get, you know, out to the street. And you realize the order is wrong. Enter Presto Automation, which is using AI to improve the drive-through process.

Can I get the Western bacon, cheeseburger? Okay, got it. Anything else? Uh, no, actually, can I actually get, can I get the California classic double cheeseburger instead? Okay, I've got it. And that'll be... What else can I get for you? And that'll be it.

I have one California classic double cheeseburger, correct? Correct. Krish Nagupta is the chairman and CEO of Presto Automation. Welcome, you're here on set with us. You guys are a public company now market cap just over $100 million and maybe the most tangible way for a lot of people to relate to. to AI. Tell us the backstory. Great to be here today. And the backstory is quite simple. You know, at the end of the day, we've been in the restaurant industry, the restaurant technology industry for a long time. Over the pandemic, we realized that drive-through and automating

drive-thrues happen to be one of the most immediately actionable applications of enterprise AI and then enter generative voice and generative AI. And that actually makes all of this even more compelling. So we think we've sort of stumbled upon this amazing, it's amazing convergence of the technology being ready just at this time and the market being ready with labor costs going up and people realizing like, wow, we should be doing this. I'm getting calls from restaurant CEOs, from franchise operators, from board members being like, how can we get this into our system now? So I guess one of the great frustrations when I interact with a voice, whether it's service, whether it's at a restaurant, whether it's a service call I make to Delta Airlines.

I'm not picking on Delta Airlines particularly. They never seem to have or be able to answer the question that I have. Will AI enable it so that the question I want answered, they can actually answer? Well, it's a really good question. So I think it's an important distinction to draw between drive-thru and customer service in more complex areas. And so in drive-thru, the nice thing is it's a relatively confined problem set, and I'm working off a menu that's particular to a brand or particular to a franchise. and so I can understand your question, and I can control it, and I can answer.

Now, it's not trivial. There are a lot of complexities in it, which is why it's an interesting problem to work on. I want no pickles and no lettuce. Yeah, and then I want a strawberry milk. Oh, no, actually, I want a chocolate milkshake, and it's like, you know, okay, I want the special of the day, but it's not the special of the day. So there's all that. Now, the interesting thing is you're talking about Delta Airlines.

I think to answer your question and make you a satisfied customer for Delta Airlines involves and requires a certain amount of emotional and cognitive intelligence, which the drive-through doesn't necessarily do, but I was actually sitting down with the CEO of a big customer or big potential customer yesterday, and he said the top 5% of order takers in the drive-thrus are actually very emotionally intelligent. So if we can now use the AI to make 100%

every single order-taker be emotionally intelligent and cognitive and accurate for 100% of orders, that's transformation. And let me just offer the observation that both offers you a viable sort of business if people want to tie up with you. you're actually able to increase revenue at some of these partners, for instance, by upselling people on the drive-through. So as we think through the implications, you know, I think yours is one of the most tangible.

That's right. No, the upsell engine is supremely important to the ROI for the customers. The labor savings are nice, but at the end of the day, if you can increase a check size or your overall revenue on a store for five, six percent, that is massive. Yeah, Krishna, off to the races. Thanks for joining us today. Thank you, guys. Hold the pickles, please. All right. Thanks very much, ma'am. All right, coming up, there's an app for that. We will highlight some AI-powered applications looking to improve our everyday lives and tasks.

And the markets with a nice rally right now. The Dow is up 339 points or a little more than 1%. We'll be right back. Well, at this stage of the AI rollout, a lot of applications are focused on improving business operations. But what if I want to use AI to make my life easier or simply more fun? So we sent Mr. Fun, Steve Kovac, to explore what apps are out there. One's actually my middle name. That's your movie.

Not my last name. Look, I am not really good at Photoshop. But this, just a few days ago, Photoshop put in their genitor of AI technology into the regular Photoshop app. Really? And I was able to. So I took a picture of you guys in the newsroom. I think we have it. In case you were curious why I took that picture of you, let's show the original first.

And then show what I was able to do just by telling Photoshop, you know, certain things. Let's go to the next one real quick. So here I told, all I did was. It looks like Snapchat. All I did here was circle Kelly's face, type in, put a bow on her head. If I had to do that myself, Kelly, I would have no clue how to do it. I'm also terrible at Photoshop. I know how to upload a photo to Photoshop, and that's about it.

And then more examples here, look, we're going to improve this photo, I think, on the next one. Yeah, we improved about 50%. Oh, look at that. How did they replace it? The dead weight just left the room. So what's really interesting here. Again, this is, if you've used Photoshop before, a Photoshop professional can do this in two seconds on their own. Google has the phone now, but like the fact that they can fill in the background and make it, you never have known.

What did you say to it? All I said, get rid of Tyler? I just said, erase him. I circled you and said erase him. I hope my boss didn't hear that. And here we go. Circle out of Snowman. This was my producer, Brianna.

She suggested we add a snowman or a zombie. I went with Snowman. This is our new co-host, I think. This is, yes, when Tyler's out sick, we're going to have the snowman fill in. But again, the idea here is it just, I said, put a snowman. next to Tyler, boom, it did it. Wow. And my point here is that a dummy like me who can barely use a mouse within Photoshop

can all the sudden use those, just tell it what they want to do. But you can use that to create misinformation and disinformation. That is the other thing. And we saw that in practice just a few days ago with that fake Pentagon photo that even sent the markets down a little bit. So really cool for someone like me who might want to explore an app like Photoshop. But this is how practically we're seeing AI news. Now it's not perfect.

You can see some blurriness in those photos. No, as soon as you showed that, I went to check Adobe shares. They are up 5.5% today. And I think maybe that's too low for what you just demonstrated. This is amazing. Again, no skills here. And look, this is not the only app. ChatGPT came out on iOS last week.

It's still sitting, I believe, at number three in the Apple App Store. The implications there are huge. Great for cheating on homework or whatever. You know, the kids are using it for these days. And then Microsoft this week, they're not really necessarily doing apps, but they're calling them plug-ins that kind of built into this chat GPT. So, for example, they have Instacart. If you ask for a recipe and it gives you all the ingredients, so you say, well, put that in my Instacart and just order it.

And it does that for you too. So a lot of interesting practical ways. But the Photoshop thing just came out a few days ago, absolutely blew my mind that I can do that. So much to play around with. Thanks for erasing me. You're welcome. Thank you. Silicon, the regulation situation.

Lawmakers are struggling to regulate social media still. and cryptocurrencies. So how about tackling AI? We'll explore the potential pitfalls of its rapid ascension and what regulators should do to try and prevent bad outcomes. We're back after this. Welcome back to this Power Lunch special. So far, we've explored many of the opportunities in AI investing, in business, and how it'll make our lives better. But now we want to examine how it could make our lives worse, the potential threats of the technology and what regulation is needed to avoid a bad outcome from this technological revolution. Joining us to talk about all of it, tech entrepreneur, Shanei Bovel, Rohe Chopra, the head of the Consumer Financial Protection

Bureau, and our very own Steve Kovac. And Steve, let's begin with you and looking at how the companies involved are trying to approach this issue. Yeah, it was all about the EU this week. So Sam Altman caused a little bit of a stir a couple of days ago because he basically said, if this EU regulation that's working its way through the process right now is past as it is, we might have to pull out of Europe. Now, that is a playbook right out of Mehta's playbook when they are facing regulation. They constantly threatened. We might have to stop offering our service. He actually ended up walking that back today in a tweet saying, oh, we welcome regulation. We're not going to leave Europe. But I think, you know, that this is telling of the mindset that a lot of

these people, especially Microsoft, especially Open AI and especially Google, Sunday, Barchar, was also in the EU. We were talking earlier about MOTS in the first segment of the show. they have the moat, these companies that those three companies I just named off, they don't want to give up that moat because they've turned it into a platform already. And by having that control, they can own the future of this. They can monetize it better. And this EU law wants them to kind of disclose some of the models that they are training these data sets on. They are vehemently against that. When they're looking for regulation, what they want is this prescriptive regulation saying AI can't be used for fill in the blank. Medical is one that gets used very often. Military is another one. So that's what they want this prescriptive regulation, not saying, you know, this broad, open

regulation. And we see Sam Altman kind of stepping in it this week about that. How are you thinking about AI as specifically it affects consumer protection, financial protection? Well, one of the things that I just don't think is true is that AI is unregulated. In fact, many of the uses of automated decision. making. The current laws apply. We have anti-discrimination laws. We have laws against deception. We have laws relate to employment, so many others. And one of the things we like to make clear to the companies we supervise is that your machine or your AI is not some exemption out of the existing

law. We see how AI is being used in customer service, in lending and originations. And we're looking about how the existing laws on the books can actually rein in some of those abuses. Yes, there are other places. We may need more rules. But I just want to make sure as an initial step that people aren't using their machine as an excuse to violate existing law. Shadneed, what are your concerns and where do you think regulators should be most focused? Yeah, my concerns kind of fall into three categories, right? We have the shorter-term risks, the disruption these systems could have to our information ecosystems. So what does democracy look like in a world where it's not just human-generating content,

but artificial intelligence at mass scale, people kind of mass-producing misinformation to disinformation. Then we move more so into the mid-term, or medium-term risk with artificial intelligence. That's the disruption to the workforce. And then there's this longer-term category where there is a little bit more controversy, things like AI alignment and control. And that part of the risk spectrum seems to get a lot of focus. But I think regulators really do need to be kind of zooming in to the short-term and the medium-term risks. And then long-term risk needs to stay in this category of exploration and research. Roy, let me turn back to something you were just saying there.

And I'm curious, in your realm, which is financial protection. let's say I'm applying for a mortgage. There seems to me to be an algorithmic method to decide whether I should get the mortgage or not. Or maybe that's artificial intelligence. Where's the line drawn between sort of simple algorithmic decision making and artificial intelligence decision making to determine whether I get the mortgage or somebody else gets the mortgage or whether I get this credit score or somebody else gets that credit.

Well, you're hitting on something which many people believe artificial intelligence is sometimes just a marketing term. You're right. There's many places where automation has been a part of industry advances for generations. We've been using credit scores, other things that are based in data and algorithms. I think really where there might be a breaking point is when it does come to generative AI. when you start having the creation of voices and other content that mimic human behavior. Look at American consumers now when dealing with banks. Many are dealing with chatbots who are given human-like names. And I think that we're going to expect that those bots or others are going to be following the law

the same as a human customer service rep. We're going to make sure that when those algorithms or AI are being used, for mortgage origination that all the existing laws on the books apply. And we've really told the industry that if they don't understand how their AI is making decisions, they may not be able to use it and comply with existing law, but it does raise this question, what really is new about AI and what's really the same?

I mean, you can easily imagine, look at crypto, where we're years into highly successful businesses and now just starting to see the regulatory hammer drops. So, Steve, they've got to be very careful. Yeah, and I would also point out what this discussion about copyright I knew you talked about earlier in the show, but that seems to be, as regulators are looking at this, one of the lowest hanging fruit that they can go after,

copyrighted material, name, image, and likeness. That seems to be something everyone agrees on. All right. All right, folks. Thank you very much. They're playing our song. Sheney, thank you, Roet. Thank you, Steve.

Thank you as well. Thank you. And thank you all for watching. This special edition of Power Lunch. The Power of AI.