Episode 45 of the Marketing AI Show with Paul Roetzer and Mike Kaput covers the business of ChatGPT, the disruption on politics that AI is having, and the reality of the growth of AI as big tech continues layoffs.
Listen or watch below—and see below for show notes and the transcript.
This episode is brought to you by MAICON, our 4th annual Marketing AI Conference. Taking place July 26-28, 2023 in Cleveland, OH. Current discounts end May 12, so register early!
00:03:41 — ChatGPT business
00:10:48 — AI and politics
00:25:34 — Big tech layoffs
00:38:22 — Runway Gen-1
00:40:09 — PwC’s big bet on generative AI
00:42:25 — AI, empathy and healthcare
00:45:51 — Replit’s funding round
00:48:08 — Hinton leaves Google
Big changes are coming to ChatGPT
OpenAI just announced two big updates to ChatGPT. The first is a soon-to-be-released subscription tier called ChatGPT Business. Designed for enterprises, the plan will follow OpenAI’s API data usage policies. That means user data won’t, by default, be used to train ChatGPT. The second is a feature that now allows ChatGPT users to turn off their chat history, which will prevent conversations from being used to train ChatGPT. What does this mean for business leaders using ChatGPT? Will enterprises who have prohibited the use of ChatGPT change their tune with these changes?
We got a startling preview of how AI is going to impact politics
In the U.S., the 2024 presidential election season kicked off with an attack ad generated 100% by artificial intelligence. The ad imagines a future dystopia where President Joe Biden remains in office after next year’s results. The images, voices, and video clips are stunningly real and were created with widely available AI tools. And they foreshadow an election season where AI can be used by all parties and actors to generate hyper-realistic synthetic content at scale.
At the same time, lawmakers in the U.S. and Europe signaled this week that they’re taking more aggressive action to regulate AI. In the U.S., four major federal agencies, including the Federal Trade Commission and the Department of Justice, released a joint statement on their stance toward AI companies. The agencies clarified that they would not treat AI companies differently from other firms when enforcing rules and regulations.
In Europe, the European Parliament has reached a deal to move forward on the world’s first “AI rulebook,” the Artificial Intelligence Act. This is a broad suite of regulations that will govern the use of AI within the European Union. These include safeguards against the misuse of these systems and rules that protect citizens from AI risks. Our team has some opinions. You’ll want to listen to this part of the podcast.
AI’s major impact on big tech companies
A recent round of tech earnings calls saw major companies like Microsoft, Google, and Meta displaying strong or better-than-expected results—and some of that growth was driven by AI.
In Microsoft’s case, Azure revenue was up 27% year-on-year and Microsoft said it was already generating new sales from its AI products.
Google was less specific about its AI plans but committed to incorporating generative AI into its products moving forward. Reports have surfaced that Meta is playing catch-up to retool its infrastructure for AI but still saw an unexpected increase in sales in the past quarter.
At the same time, these companies face enormous pressure from shareholders to get leaner. Some have conducted layoffs already, with more expected to come. And they’re all relying on AI to capture efficiencies. We saw a stark example of this in practice with a recent announcement from Dropbox that they’re cutting staff by 16%, or 500 people. How should knowledge workers think about this? What steps should we be taking?
Today’s rapid-fire topics include Runway Gen-a for mobile, PwC invests $1 billion in generative AI, and AI and human empathy in healthcare, Replit’s funding round, and Hinton’s Google exit.
Listen to this week’s longer-than-usual episode on your favorite podcast player and be sure to explore the links below for more thoughts and perspectives on these important topics.
Disclaimer: This transcription was written by AI, thanks to Descript, and has not been edited for content.
[00:00:00] Paul Roetzer: Welcome to the Marketing AI Show, the podcast that helps your business grow smarter by making artificial intelligence approachable and actionable. You'll hear from top authors, entrepreneurs, researchers, and executives as they share case studies, strategies, and technologies that have the power to transform your business and your career.
[00:00:20] Paul Roetzer: My name is Paul Roetzer. I'm the founder of Marketing AI Institute, and I'm your host.
[00:00:28] Paul Roetzer: Welcome to episode 45 of the Marketing AI Show. I'm your host, Paul Roetzer, along with my co-host as always Mike Kaput, Chief Content Officer at Marketing AI Institute and co-author of our book, marketing Artificial Intelligence, AI Marketing in the Future Business. This is speaking circuit week for us.
[00:00:47] Paul Roetzer: I am actually coming to you from Austin, Texas, a beautiful resort in Austin, Texas, where I'm talking about, transforming the senior living industry with artificial intelligence. I also have a talk, I think I'm on Friday in Charleston, South Carolina. Couple of virtual things too. And then Mike, you're at the, creator.
[00:01:09] Paul Roetzer: Economy Expo tomorrow in Tuesday in Cleveland, is that right?
[00:01:13] Mike Kaput: Yeah. Wednesday is my talk, but I'll be okay. Lurking around there throughout the week as I do other stuff. And then I leave Thursday for beautiful Punana to speak to a network of agencies.
[00:01:26] Paul Roetzer: Taking one for the team there. Huh? Gotta go to pun for a week.
[00:01:29] Paul Roetzer: It's tough. Yeah. Yeah. It's a, there's a, it's funny, the, you know, obviously the interest in AI has, exploded amd so Mike and I have been keeping pretty busy with the speaking circuit. There's a week in June, I don't know, we're going to do the, podcast that week. We might have to record. But Mike is in Chile.
[00:01:48] Paul Roetzer: And I'm in Romania and Italy. All within like a three day span. I think it's, so that'll be an interesting week for the podcast, but we'll figure it out. Maybe. We'll, we'll do a bathtub or something like that.
[00:02:00] Mike Kaput: Yeah, that's one where you need to start getting virtual avatars. For us to do it more
[00:02:05] Paul Roetzer: traveling, that's not bad.
[00:02:07] Paul Roetzer: We could explore that. The synthetic versions. Interesting. All right, well let's get into this one. So this episode is brought to us by the Marketing Artificial Intelligence Conference, Macon, which is coming back to Cleveland July 26th. The 28th. We just released the, preliminary agenda last week.
[00:02:26] Paul Roetzer: It's about 80% or so of the agenda is on there. Still have a lot of. General session and keynote announcements to be made over the next month or two. So stay tuned for those. But you can actually go up to Macon ai so it's ai, and check out some of the speakers we've got lined up already, and get in early.
[00:02:44] Paul Roetzer: I think May 12th is the next price increase. So if you're hoping to join us in Cleveland, Try and take advantage of that early bird pricing before May 12th. So again, it's Mayon, M A I C O N ai. I hope to see you there. And again, if you're new to the show, we pick three big topics for the week. It is increasingly hard by the week to do this.
[00:03:05] Paul Roetzer: I feel like. We record every Monday. We record last Monday, and I think by Monday afternoon we had like three new topics already, like before the other show didn't even come out yet. So we pick three topics, we go through those and then rapid fire. And today we've got about five rapid fire items. We got a lot to cover.
[00:03:21] Paul Roetzer: Yep. A little bit of time to do it. So I'm going to turn it over to Mike and let's, let's get going with the first topic, Mike. Thanks
[00:03:26] Mike Kaput: Paul. First up, we've got big changes. That are coming to ChatGPT. So OpenAI just announced two big updates. Two ChatGPT. So the first is going to be a soon to be released subscription tier.
[00:03:42] Mike Kaput: They're calling ChatGPT Business, and this is designed for enterprises. The plan is going to follow OpenAI's API data usage policy. So what that means is if you're on this tier, The user data that you give ChatGPT to produce prompts will by default not be used to train ChatGPT. So for anyone who doesn't know today, unless you are specifically turning off certain settings through the API, anytime you share content with ChatGPT or ask it to produce outputs for you, you are also helping train the model The second.
[00:04:20] Mike Kaput: Is a feature that now allows you to start preventing this as an individual user as well, so we can now, we are now going to start seeing individual users at any tier have the ability to turn off their chat history. When you do that, anything that you. Give ChatGPT in terms of data or conversation will not be explicitly used to train, ChatGPT.
[00:04:43] Mike Kaput: So these kinds of, privacy and security features are definitely more appealing to some of the concerns out there about how OpenAI trains the model and as well as the concerns that. Enterprises have when trying to use these tools in a compliant way. So first up, Paul, I wanted to ask you, what do these changes mean for business leaders who are using ChatGPT or thinking about using it moving forward?
[00:05:10] Paul Roetzer: I just think it's interesting that OpenAI, you know, is making the play. It seemed really obvious this was the direction they were going. They know that that. You know, enterprises are not going to want their data in future training sets for the next foundational models. So it was sort of inevitable that they were going to do this.
[00:05:26] Paul Roetzer: The conversations we've been having lately with enterprises and Is this exact concern. Like everybody's trying to figure out the play with large language models. Should they be, you know, customizing and training their own? And if so, which language model company or which AI application company should they be working with?
[00:05:41] Paul Roetzer: And how does that work? And this is sort of like, Again, almost every conversation I'm having at the C-suite level, it's coming back to this, what, how do we build a strategy around our use of large language models? And obviously inherent within that is we are not going to give our data to anybody. There was a crazy example.
[00:06:00] Paul Roetzer: I think you, you and I traded this back and forth maybe, but that guy on Twitter, we'll find the link and put it in here. Who like, Patched in G PT four to his bank accounts and credit cards, like gave it all its personal information, investing details so that it could help him find savings, like find anomalies and expenses and his credit cards and all these recurring costs.
[00:06:22] Paul Roetzer: Cause it's like so super utility, like great function. But in the process, giving all of his personal financial data to OpenAI, which looks like crazy. And yet, you know, I know that a lot of people don't realize. That's what's happening when they're putting their data into there. I was actually, I've had just in the last week I had conversations with, some attorneys who weren't aware of that.
[00:06:46] Paul Roetzer: And some other industries where they may have people who are using these tools unbeknownst to the corporation and could easily be putting in meeting notes for summarization or things like that where there's confidential information contained within them. Asking it to summarize PDFs where you're cutting and pasting stuff into there, soap.
[00:07:05] Paul Roetzer: There's all these uses, and I think the key takeaway for me is that, Again, if you're not aware, you have to be careful what information you put into these tools. You have to understand the usage of your data, how they're going to apply it, and to make sure it remains your confidential, proprietary data.
[00:07:24] Paul Roetzer: And I just think this is going to be something, a continued theme we'll talk about on this show because almost every enterprise I'm aware of is starting to realize they have to solve for this and trying to figure out what do they do, who do they talk to? How do they move forward? So, yeah, I don't, and as a, as someone who uses chat g ppp, the one thing I thought of was like, why do I have to turn off my, like, chat history like that?
[00:07:49] Paul Roetzer: That's a really useful function for me is to like, see the things I've function. I have like, you know, dozens of these things I've done U QB T four, to create and like to find them is kind of hard right now. So I'm all constantly scrolling, but I like having that history there. I don't want to have to give that up just to like, not have my data, but at the same time, that's probably the barrier.
[00:08:12] Paul Roetzer: They want. They, they want your data so they don't want you to easily turn it off. So, yeah, it just seems like the direction it's all going to go. And something for people to keep an eye on. So
[00:08:22] Mike Kaput: given that, do you see this as leading to more enterprise adoption or usage of ChatGPT? Is it too early to tell?
[00:08:32] Paul Roetzer: I would guess it at least puts them in the conversation. Again, if you're working in a, an industry like insurance or financial services or healthcare or, you know, I mean, honestly, really any enterprise that protects its data and has a CIO and has security protocols, then governance, and you, you're going to ask these questions regardless.
[00:08:53] Paul Roetzer: Previously, OpenAI just may not have been. A, a, a platform you would consider if they didn't have these protocols in place. So I think it's almost just like essential for them to be part of the conversation around corporate language AI
[00:09:06] Mike Kaput: strategies. Gotcha. And also I think it's important to layer in some context here, these moves, they're changes, these aren't happening in a vacuum.
[00:09:17] Mike Kaput: OpenAI has faced some pretty serious legal and regulatory challenges, notably with Italy. Trying to ban chat, G B T and the eu. Looking more closely at how OpenAI trains models, how does this move or these developments relate to what they're dealing with on a legal and regulatory
[00:09:37] Paul Roetzer: front? I assume it's all connected.
[00:09:39] Paul Roetzer: Just last week, didn't Italy, reinstate ChatGPT? So I assume there was some sort of concessions made on OpenAI's part that would enable them to get back in amd try and proactively prevent other countries, in the EU from, you know, Shutting it off. So yeah, it all seems to be connected. We know OpenAI from past episodes.
[00:10:02] Paul Roetzer: We've talked about, we know OpenAI is in conversations with the US government, you know, on oversight and regulations and, trying to be proactive there and I'm sure doing some, To kinda protect the innovation side of this. So yeah, I think it's all connected, which is why it's probably going to be an ongoing story on this podcast.
[00:10:23] Paul Roetzer: At least. It's just, it touches every aspect of society and business right now.
[00:10:28] Mike Kaput: Yeah. And speaking of how it is starting to affect society, don't, be upset with me if I am sighing during this next topic. It's not because I'm not interested in our podcast, it's because what we have feared would, would come to pass, has started to, because we just got a startling preview of how AI is going to reshape politics.
[00:10:51] Mike Kaput: So in the US here, we're kicking off a 2024 presidential election season. Our election seasons are painfully long, and this one kicked off with an attack ad. That was 100% generated by artificial intelligence. And basically it's a video that imagines like a future dystopia where President Joe Biden has remained in office after next year's results and all this terrible stuff happens.
[00:11:15] Mike Kaput: So, obviously his, Republican focus, Produced video and the images, the voices, the video clips, all AI generated are really, really convincing. And they were created with widely available AI technology, and they kind of foreshadow this election season where AI is going to be used by all parties in all actors, on every part of the political spectrum to generate hyperrealistic synthetic content at scale.
[00:11:44] Mike Kaput: And we'll talk about that a little bit here because at the same time, I think because of threats like these and what else is going on, governments do seem like they are. Speeding up taking some type of action around ai. So lawmakers in the US and Europe both signaled this week that they're taking more aggressive action to regulate ai.
[00:12:06] Mike Kaput: For instance, in the US we saw four major federal agencies, including the Federal Trade Commission and the Department of Justice release a joint statement on their stance towards AI companies, and they made it very clear they would not treat AI companies. Different from other firms when enforcing rules and regulations.
[00:12:26] Mike Kaput: So they've come out with quite a strong stance saying, look, if you break the rules, you will face consequences regardless of what technology you're developing now in Europe. European Parliament reached a deal to move forward on the world's first, you know, kind of comprehensive AI rule book. It's called the Artificial Intelligence Act, and it's been, around for probably almost a year now, or hypothesized and debated over in the European Parliament, and it's a broad suite of regulations.
[00:12:55] Mike Kaput: That govern the use of AI within the European Union. So there's tons of safeguards in there that protect against things like misusing the systems and having rules that protect citizens from AI risks. So there's a lot going on when it comes to how AI is impacting governments and society and how those forces are reacting.
[00:13:16] Mike Kaput: So Paul, I know you had a pretty strong reaction to the political ad that was released this week. Can you walk us through. Your thoughts. Specifically you had worn that despite how much of a hot topic AI is, that the average citizen still really has no idea that AI can create realistic synthetic media at a level of accuracy or scale that we're seeing
[00:13:39] Paul Roetzer: today.
[00:13:41] Paul Roetzer: Yeah, the thing that I took away, the ad itself isn't anything crazy. Like if you've been following along in this space, it's not like there was some breakthrough in ai, you know, technology that generated some kind of creative output we've never seen before. It was basically a mashup of, AI generated images that strung together with some text and some audio, but it, the thing that, Really was a triggering point for me is it's just the prelude to what's about to happen.
[00:14:11] Paul Roetzer: So I mean, we're going to see. AI creating content at a scale we've never seen before. Personalization of ads, personalization of messaging, text messaging, emails, video content, like all of it, all the best of what's available. The most advanced technology is going to be used in these campaigns. Like there's, it's not even a debate.
[00:14:35] Paul Roetzer: And then the challenge, you know, somebody said on, on my LinkedIn post was about like, well, when lawmakers step in, it's like, They're the people that are using the technology, what are they going to do? Stop themselves from using it? So I feel like the significance is really the fact that it's just a sign of things to come.
[00:14:54] Paul Roetzer: I, you know, one person was commenting. Oh, it's just like, you know, who cares? It's like Photoshop. Like we could all always, always do this with Photoshop and it's like, I think that's just missing a point. Like, like to do something like this with Photoshop would've required expertise in Photoshop to do something like this.
[00:15:11] Paul Roetzer: And Mid Journey doesn't require the depth of expertise. Sure, you'd need to know how to prompt it and stuff, but the average person can do this. And so the average, army of people that are working for these campaigns, Can create almost endless amounts of this content. And so that's the part that really worries me is the ability to create this stuff at scale, target it at an individual level basically.
[00:15:36] Paul Roetzer: I mean these political campaigns are some of the most advanced campaigns we've ever seen. Most marketers probably aren't even aware how advanced the stuff, these political campaigns too. They have hundreds of millions of dollars, not billions of dollars to put into these campaigns. And so I just feel like we're about to see.
[00:15:53] Paul Roetzer: The best and worst of what AI can do all at once for the next, what have we got? 16 months, 17 months, whenever this election happens. Yeah. So we are about to get, firsthand look at what this AI is capable of. And as you highlighted, my biggest concern is that the average US citizen has no idea AI can do this though.
[00:16:16] Paul Roetzer: Like they're not, they're not going to know the difference. And that's, that's a problem. Like, there, there's a, a need for, very rapid education because, I mean, political campaigns historically they do disinformation, misinformation, propaganda, but by nature what politics is. But never at a scale like this, never with the ability to manipulate people, their emotions and their behavior the way that they'll be able to now.
[00:16:42] Paul Roetzer: And that's what worries me about it.
[00:16:46] Mike Kaput: So it doesn't really seem like we're in a time where brands and businesses can totally stay out of politics. So like what, as a professional or a business leader, seeing where this is going and where AI generate content will soon be a norm in terms of what we are bombarded with and have to respond to as business people.
[00:17:09] Mike Kaput: Like what, what should business people be thinking about or doing? To prepare for this.
[00:17:17] Paul Roetzer: Honestly, at this point, I think it's doing their part to help educate society. Like I, I don't know how else to say it. Like, I mean, you'll get. You're going to start getting questions from family members and friends. It's going to be a topic of conversation.
[00:17:31] Paul Roetzer: Like this is going to start surfacing. And people that would never talk to you about AI are going to start asking you like, do you know how they did that deep fake thing? Or do you know what they're talking about with all this misinformation? Like, and you're going to be the one that needs to bring reason and understanding to the conversation.
[00:17:48] Paul Roetzer: . So I think that if nothing else, it's just preparing yourself to help educate. Within your own company, but you know, larger point within society, within your family, within your friends, just about what's happening because it's going to be really hard to know what to believe. And I do think that in some ways, like sources of truth are going to be really, really important.
[00:18:15] Paul Roetzer: Like if your, your family. Or coworkers are the kind of people who tend to believe what they see online on Facebook, on Twitter, wherever they're getting their information. TikTok, there has to just be an increased awareness that what they're seeing may not be real and true. Now, again, this isn't new to society.
[00:18:35] Paul Roetzer: We've been dealing with this for a long time. But again, it's just going to be the scale and I really think we're going to need to just find. The the sources of information that we know we can trust and if we see something verify like it, I don't, I don't know, like I don't know that we can rewire society in the next 18 months of how to process this kind of information, but I really just think it's doing your part to help educate as best as possible.
[00:19:00] Paul Roetzer: Without alienating, like it, again, it's politics. Like, it's really hard to have these conversations without . Someone getting mad or someone, you know, throwing dirt at one side or the other. And we're just trying to kind of raise awareness of it as a, as a content and an inter information vehicle, and the need for people to understand how this stuff works and what it's able to create.
[00:19:20] Paul Roetzer: I just, I don't know what else to do at this point, honestly. That's why we. Put it up on LinkedIn and why we wanted to throw it into the podcast is just make sure that the conversation's happening in some small way. Yeah.
[00:19:32] Mike Kaput: So let's talk for a second about these varying government responses to AI technology out there in the us.
[00:19:41] Mike Kaput: It seems like, from what we've heard from the federal agencies that I just talked about that we're moving towards. Essentially regulating AI by applying existing laws to AI companies, whether they're, you know, that apply to non-AI companies as well. Whereas it seems like the EU is attempting to craft AI specific legislation.
[00:20:03] Mike Kaput: How do you view kind of the differing approaches here? Is there one that's better than the other
[00:20:07] Paul Roetzer: or. I don't know, like, you know, again, I think this is an important topic to surface. We've talked previously about the challenges, the AI act within the eu and, you know, the technical challenges of o overseeing that, kinda governing once they have it in place.
[00:20:23] Paul Roetzer: And we've talked about kind of the, what appears to be lack of action from the US government. . So I think it's an important topic. There was a fast company article that will drop in the show notes, that, you know, they kind of brought up the point of. That this, what you were saying, the US is going to try and kind of apply these existing laws where the EU is trying to get specific.
[00:20:43] Paul Roetzer: And the challenge they said is like, it's going to get really hard to, to apply it to specific technologies, especially as this stuff scales and all these new technologies keep popping up and I have to keep applying new laws. So it seems like, a kind of more general guidance. Might be the better path here.
[00:21:03] Paul Roetzer: . But you know, when I read the joint statement, so the joint statement was from the Consumer Financial Protection Bureau, department of Justice, civil Rights Division, the Equal Employment Opportunity Commission, and the Federal Trade Commission. When I read it, I just felt like it was a lot of, okay, like, how are you actually going to govern this though?
[00:21:23] Paul Roetzer: . Like I, it just seems like. A blanket statement to just say, we're looking at this and we're aware we're not going to let people get away with stuff. With no real clear guidance of what that actually means or what actions they would take. So I can give you an example. So in the joint statement, they say that these federal agencies are responsible for enforcing civil rights, non-discrimination, fair competition, consumer protection.
[00:21:48] Paul Roetzer: So totally get this, like they don't want AI used in determining, your, your ability to get home loans or jobs or like these things that are obviously like, okay, that makes sense and FTC doesn't want you to do false advertising. Sure. Okay. Like, I can see how these existing laws can govern that. .
[00:22:05] Paul Roetzer: But it really starts to get kind of messy when you start looking at some of the specific areas. So like they said, the FTC Act, Doesn't want to be able to use, discriminatory impacts or to make claims about AI that are not substantiated or deploy AI before taking steps to assess and mitigate risks.
[00:22:23] Paul Roetzer: Finally, the FTC has required firms to destroy algorithms or other work product that were trained on data that should not have been collected. Okay, so if we just take that paragraph, do not deploy before taking steps to assess and mitigate risks. So does this mean that the FTC act is cool with how GT four came to market?
[00:22:43] Paul Roetzer: Like, do they, or mid journey or stable diffusion, like they're, are they looking at those things because th those would all seem to fit under this umbrella, in which case then they should give more guidance. That says specifically, here are some examples of how this has been applied. . So I look at this and say, you're basically saying, we're going to do this because you haven't apparently done it, or at least you haven't shared with.
[00:23:07] Paul Roetzer: Like the general public, how you applied those and then has required firms to destroy algorithms or other work product that were trained on data that should not have been collected. Like what? Like copyrighted materials that trained image generation technology. So they have to destroy stable diffusion.
[00:23:24] Paul Roetzer: Like, so that's where I'm saying like it's this sort like sort of broad general statement that doesn't actually appear to have. Anything tactical behind it. . And, you know, and the, you can again go read the joint statement yourself. It's only three pages, but they get into three specific areas that, again, just sort of called to me like this is just this general guidance.
[00:23:45] Paul Roetzer: So they have data and data sets. So automated systems that can be skewed by unrepresented or imbalanced data sets. They have model opacity and access. Talking about these, these things being black boxes. That is what Genive AI is like. Did. They don't know why it's generating the words it's generating, like, so all of this applies to this.
[00:24:06] Paul Roetzer: . This lack of transparency often makes it all the more difficult for developers, businesses, and individuals to know whether an automated system is fair and then design and use, which is developers do not always understand or account for the context in which private and public entities will use their automated systems.
[00:24:22] Paul Roetzer: So I, I get 'em. My overall take here is, I think it's good that the federal government is saying something. I don't feel like this is anywhere near as specific as the copyright office guidance from March 16th. I think to the US copyright office's credit, they were very specific in their guidance.
[00:24:41] Paul Roetzer: . Whether it's right or wrong, or whether it's going to have to evolve to allow some form of AI authorship, that's to be debated. But they put out very specific guidance and then they, they announced listening sessions to hear feedback. This to me is like, someone told these agencies, they better be doing something and they teamed up and put out a statement through the PR team.
[00:25:04] Paul Roetzer: And if you press them on what does this actually mean? They probably have no idea. I hope I'm wrong, but that's Sure. What this seems like as a former PR person, this is a lot of like, Hey, we're, we're working on this and thinking about it. Gotcha.
[00:25:20] Mike Kaput: So our third kind of big topic pulls together a few threads that we're seeing and where we're getting to is that we're seeing that AI is having a really major impact on the profits, earnings, and business health of big tech companies.
[00:25:37] Mike Kaput: But it's not always the impact that maybe employees would like. And so here's a couple of data points. To justify why I say that. We had a recent round of tech earnings calls and we saw some major companies like Microsoft, Google, and Meta displaying pretty strong or better than expected results given the economic environment and.
[00:25:59] Mike Kaput: Some of this growth was explicitly driven by ai. I mean, the companies are talking about AI far, far more on earnings calls than they were a year ago. In Microsoft's case, Azure revenue was up 27% year on year, and they actually said they're already generating new sales from the AI products they've released.
[00:26:19] Mike Kaput: Google was a little less specific about its AI plans, but it committed to incorporating generative AI into its products moving forward. And we have seen some reports that Meta is playing a lot of catch up to retool its infrastructure for ai. Kind of getting caught flatfooted maybe with their emphasis previously on the Metaverse, but they still saw an unexpected increase in sales in the past quarter.
[00:26:45] Mike Kaput: And then amidst this, there's another side to this coin. All of these companies face enormous pressure from shareholders to get leaner. Some have conducted layoffs already and there are some more that are expected to come, and they're pretty clear on all of them saying they're relying on AI to evolve their businesses and or capture efficiencies.
[00:27:09] Mike Kaput: And so this really like. I think we danced around the topic for a while, but it really kind of came home to roost. In a very stark example with a recent announcement from Dropbox that they're cutting staff by 16%, that's about 500 people, and they explicitly said one of the major reasons for the cuts is artificial intelligence.
[00:27:30] Mike Kaput: So co-founder and CEO Drew Houston wrote. In a letter to staff that Dropbox needed to act with urgency to seize the opportunity presented by artificial intelligence. But to do that, the company needed a different mix of skills. Than it has today. He said, in an ideal world, we'd simply shift people from one team to another, and we've done that wherever possible.
[00:27:58] Mike Kaput: However, our next stage of growth requires a different mix of skill sets, particularly in AI and early stage product development. So I wanted to kick things off here amidst, you know, better than expected earnings companies getting leaner, but we're still seeing layoffs and AI powered. Business evolution.
[00:28:19] Mike Kaput: Can you talk to us a bit about the pressure that major companies are facing in boardrooms, executive meetings to adopt ai, to get leaner, to use it, to get more efficient? What have you seen and heard so far in
[00:28:31] Paul Roetzer: that conversation? There seems to be a, a growing awareness of AI's potential. To drive much larger efficiencies in organizations.
[00:28:45] Paul Roetzer: And as we talked about in the knowledge work episode, efficiency can sometimes be code word for reduction of workforce. So I do feel that given conversations I've been having with some bigger companies, that there is going to be, increasing pressure. To find ways to be as efficient as possible for a lot of different reasons as we previously covered it, length in the knowledge work, episode.
[00:29:17] Paul Roetzer: But, I'm Optim, I want to say I understand there's a lot of like, big challenges right now in ai. There's topics we talk about that aren't always optimistic and hopeful. And I'm continuing to like, Think about what are, what are the opportunities here? amd, the Dropbox example doesn't help cause they're just like, we're just going to get rid of these 500 people and Yeah.
[00:29:44] Paul Roetzer: Find people who can work on ai basically. I don't think theirs is specifically saying, we're going to replace these people with ai. Theirs is saying, we got a lot of people who aren't contributing toward our advancement in ai and we, we fell behind. So Dropbox, he said like, we've been working on AI for years, but like we.
[00:30:01] Paul Roetzer: We weren't where we needed to be, and we need people on board who can get us where we need to be. So I would say, I'm optimistic that in a lot of industries there's going to be time to figure out how to redistribute these workforces. . And there's going to be, an effort to find roles and opportunities to increase output to, to, to really do more with the, the p time save.
[00:30:33] Paul Roetzer: I don't think SAS is one of them. Like this is the tech industry is, is going to be brutal. Like that, there's just no way around this. . I think there's going to be lots of middle management that are deemed unnecessary. I think there's going to be lots of developers who don't know how to build AI tools that are going to be expendable.
[00:30:58] Paul Roetzer: And I just think it's going to happen so fast because the pressure on the SaaS companies in particular is so, is so great. And I have talked with quite a number of SaaS companies in this space. So I do think that the Dropbox example could, could happen. I think there's a lot of product teams at these SaaS companies who have no idea what they're doing with ai.
[00:31:18] Paul Roetzer: . Couldn't build a product roadmap for AI because they, they have no background in it and they're scrambling to try and figure it out on the fly and they don't think, they don't go deep enough to actually. Realize the disruptive force it's going to have on their, their product roadmap. So I think SaaS, we're going to keep seeing these stories just one after another.
[00:31:35] Paul Roetzer: You know, we had Meadow as the big example of 10,000, you know, being laid off. But thing this is going to keep happening. The other thing I think that's happening is these, these big companies like Google and aws or Amazon AWS and Microsoft Azure, like they're trying to find markets for their technology.
[00:31:53] Paul Roetzer: . So if you think about. You know, like a Google or a Microsoft, or again, Amazon, for 20 years they've been working on ai, but they just didn't productize it. Like the, it was, it was mostly internal technology, largely. And so now you're faced with this, you know, reality of, wow, we gotta turn this into products.
[00:32:11] Paul Roetzer: Like, AWS as an example. They have, you know, in our book we highlighted in chapter one, they had like, what, like 30 or more pre-trained AI models. Like, so this was, we wrote our book in. Early, late 2021, early 2022. And so we highlighted all these pre-trained models that live in these clouds, but none of them were things you or I could go use.
[00:32:31] Paul Roetzer: Yeah. So like if we wanted to actually take AWS's personalization or content extraction or image recognition, you needed developers to help you build these things. And all of a sudden fall of 2022 hit and. Chat and all of a sudden, like, we can just use these tools, like we don't need developer friends.
[00:32:52] Paul Roetzer: And so I think that's what's, there's just so many variables happening, but those are a couple of the bigger trends that are going to impact this. So yeah, there, there's a lot of pressure though. I, again, I'm, I want to make sure we inter, we. Senses of hope and optimism into these conversations. Yeah, and I do really think that there are going to be a lot of industries, maybe ones that are dealing with shortages of workers, but I'm thinking of even the senior learning industries.
[00:33:18] Paul Roetzer: I'm sitting here in Austin, Texas, they have a massive shortage of workers. You know, maybe, maybe AI is a help to that. amd so I want to find the silver lining wherever we can. And I think there's going to be a lot of it. It's just right now that these negative things are sort of dominating the headlines.
[00:33:35] Paul Roetzer: Well, in
[00:33:36] Mike Kaput: terms of that silver lining, it might be worth spending a moment on. If you're kind of new to this topic, you might be sitting here thinking like, well, I don't really know anything about artificial intelligence. I only know about my area of expertise, and it's like I. To your point, that's a big opportunity because you don't need a PhD in AI to take what you're very good at and start applying AI to it and also understanding how AI is going to affect it.
[00:34:04] Mike Kaput: You have an opportunity potentially to be that person in your organization that sees where your domain of expertise is going. In a world that's AI first
[00:34:15] Paul Roetzer: hundred percent, and actually that's, that's like a universal theme that I'll say when I'm. Everyone is, sort of overwhelmed by this topic. It's everyone's uncertain.
[00:34:27] Paul Roetzer: Some people are very afraid of it and just want to ignore it. And you do have the chance to be someone in your organization who raises their hand. And so I want to, I want to figure this out. . So we, you know, in the Knowledge work episode, we talked about this idea of forming an internal AI council, using that council to build responsible AI principles, genive AI policies, figure out the impact on the organization, be the person that starts that.
[00:34:51] Paul Roetzer: I don't care if you're 22 or. Or 72, like raise your hand and say, I think we need to be proactive here. Let's figure what it means to our company, our workforce, and let's start planning now. Let's not wait till it hits our industry. And so I think that's the opportunity again, if you're listening to this podcast, I get that there's a lot of.
[00:35:10] Paul Roetzer: Topics, and it's kind of heavy and it's not always super optimistic, but you can turn it into something optimistic if you take all this information and just move forward in your career and say, no, no, I'm going to, I'm going to figure this out and I'm going to help my team figure this out. So yeah, I think the best way to solve for this is to just be proactive with the knowledge you're gaining and do everything you can to, to achieve a positive outcome.
[00:35:34] Paul Roetzer: I mean, that's what we're doing every day. It's like, yeah, we could easily get. Discouraged by all of this news, but for me and Mike, it's just like, no. Like let's get this information as many people as we can, and let's do everything we can to make sure it's done responsibly. And in the end, like we can go to bed at night thinking we've tried to make an effort to have a positive impact today on the industry, on society, whatever.
[00:35:56] Paul Roetzer: And so I think if you do the same, you'll find hope in the information.
[00:36:00] Mike Kaput: I think that's really good guidance. If I'm starting to think as a knowledge worker of how to start evolving in terms of like a couple just really simple practical steps. Do you have any recommendations, if I have little or no AI experience, like what should I be doing Right that, you know, this week?
[00:36:18] Paul Roetzer: Yeah. The thing you and I talk about all the time is understanding is the first step. Always like you, you just have to develop a. Confidence in the topic to be able to take those proactive actions. So that means taking an online course or reading a book, or going to a conference, like whatever it is. However you best learn, consuming the podcast.
[00:36:39] Paul Roetzer: However you like to learn, do that. Go deep on it. . And you can do it in a week or two. Like you don't, I'm not saying like, spend the next two years on this, just get to a point where you're confident enough. And I've seen it like I've done talk for corporations and corporations. I'm in an organization with a thousand marketers, like who am I at 25 to be the one to do this?
[00:37:00] Paul Roetzer: But no one else is doing it. And I'll give 'em guidance like, okay, here's what I would do. And you, so you can see the people starting to take the initiative to learn the stuff and get confident enough in it that they can then go and take action. And sometimes they'll reach out to me like a quick LinkedIn message because I feel like they're looking for.
[00:37:17] Paul Roetzer: Validation that's like, no, you can do this. Like no one else is going to be the one. And I just like, maybe take, if you're, if you fit into that where you're sitting in these bigger companies and you're wondering like, could I really be the person? Yes. And take this as like my one to, pep talk. You absolutely can be the person to do it.
[00:37:35] Paul Roetzer: I had one of our listeners recently took an episode, one of our episodes, and then one source from the Washington Post and took it to actually activate change within an education system. . And I was like, that was enough. Like she just, she needed something to go do. And I just, I love hearing stories like that cause it gives me hope that people are just going to be proactive and find ways to positively affect this stuff.
[00:37:59] Mike Kaput: That's awesome. So, Let's dive into a few RapidFire topics before we wrap this up here. And the first one is we just saw Runway ml, which we've talked about many times on the podcast. A major leader in ai. They create a popular AI creative suite where you can do everything from edit and create video images, et cetera.
[00:38:21] Mike Kaput: They actually just released their Gen One video to video generative AI model for mobile through the runway iOS app. So you can literally now in your pocket have these stunning AI powered capabilities when it comes to creative. What did you think of
[00:38:40] Paul Roetzer: this announcement? It's pretty cool. I downloaded it.
[00:38:44] Paul Roetzer: You can create up to five second videos. They've got some template themes that you can apply to any. Like images or videos in your library. So if you give it access to your, your photos and videos, and then you can just prompt it to create something. So, or that'll in the gen two, right now's video. Yeah.
[00:39:04] Paul Roetzer: I mean, There are, so one of my issues with adoption of gender AI is having to be in Discord to do this. Like I'm not a Discord user, so I don't use Mid Journey, and I don't, I just don't want to have to go there. I get that that's where people are doing it. So I think what Runway's doing is brilliant because it's getting rid of that barrier.
[00:39:22] Paul Roetzer: Like, why do I have to go into Discord to use something? I just, just download an app or go to the site. And so I think their adoption of runway will, you know, probably. Skyrocket. But yeah, it's really cool. Again, it's just experimental. Like creating a five second video isn't life changing. You're not going to, you know, create your next trailer for your, your demo for your company with this stuff, but you get a sense of where it's going and what's going to be possible.
[00:39:46] Paul Roetzer: So definitely download it, check it out. And again, they're not a sponsor, like it's just awesome tech that we use all the time.
[00:39:53] Mike Kaput: So next up we saw that a major accounting front, Pricewaterhouse Coopers has said it plans to invest 1 billion over the next three years, specifically to transform its operations with generative AI technology.
[00:40:10] Mike Kaput: And what's really interesting about the announcement is they say they plan to use AI to quote, automate aspects of tax, audit and consulting services. They're also going to be using this funding to recruit more AI workers, train existing staff in AI, and potentially acquire some AI companies, and they're doing this all in partnership with Microsoft and OpenAI.
[00:40:34] Mike Kaput: This seems like a pretty big swing in that industry. What did you think?
[00:40:39] Paul Roetzer: Yeah, they're not alone. I mean, the article says other accounting firms, including kpmg Ernst are also investing in ai TurboTax owner Intuit. Is building its own generative AI language model for financial management. Trained on years of interactions with business customers, like just the continuing investment here.
[00:40:58] Paul Roetzer: And I just, I don't see any end of this. I feel like we're really just at the beginning stages of this evolution. You're see it major consulting firms, we're already seeing it. Major consulting firms. The part about the talent. Interesting. Like where are you going to find this talent? Yeah. Is my big question.
[00:41:14] Paul Roetzer: So I do think that, and again, find finding hope in all of this. Every company is going to need workers who understand AI and can guide it, you know, human and machine. the human prompts. The machine knows what to do with the output. It can evaluate the output. Those people don't exist. Like they're nowhere.
[00:41:36] Paul Roetzer: So, I mean, we've thought about how do you scale up? Like if you wanted to start a company and hire a hundred people who knew this stuff. Where are you going to find them? So I think that every organization is going to, you know, that's serious about really becoming AI emergent, is going to be very aggressive in building internal education programs to evolve their teams to be, high comprehension levels with AI so they can work with tools.
[00:42:05] Paul Roetzer: Excellent.
[00:42:06] Mike Kaput: This other, this next one was pretty fascinating to me. I know it caught your eye as well. A new study just came out that looked at whether or not an AI chatbot could provide healthcare answers to patients in a way that was as accurate and empathetic. As a human physician. So to test this, basically the researchers tested both AI and physician written answers on a social media health forand they actually found out, at least in their study, that the answer was yes.
[00:42:37] Mike Kaput: The AI response is, we're actually preferred over the human physician wants, and they were rated significantly higher for both quality and empathy. What did, what was your take on this study when you started reading
[00:42:51] Paul Roetzer: about it? Two thoughts here. One, we are not by any means saying that AI is going to replace physicians because as we have said, many, many times, large language models make stuff up all the time.
[00:43:03] Paul Roetzer: . So it's not like we are at a point where we're going to flip switch and AI chat bots are going to replace the need to with physicians. Physicians likely still need to be in the loop at all times when it comes to this stuff. So, kinda a, a disclaimer here. The second is, it's one of the moments where I step back and realize that.
[00:43:20] Paul Roetzer: We talk about empathy being one of those uniquely human things. And then I started thinking about my life and realizing how horrible many humans are at being em empathetic. Like there's a lot of, I don't know if unempathetic is the the right word here, but there's a lot of people who don't have a lot of empathy Yeah.
[00:43:40] Paul Roetzer: And can't even pretend to do it. They just don't. And the AI can do it. Every time it can synthesize or simulate empathy. So what I could see is, again, this human plus machine scenario, there's lots of doctors I've had who just have no bedside manner. . Like, they're just terrible at it. And I could see an AI agent that's like, make this more empathetic is like a, a feature in the language model.
[00:44:10] Paul Roetzer: So it's like, okay, I'm, I'm responding to this patient via email. Or private message, whatever. And then there would be a add empathy button and it would almost simulate what an empathetic human would say. So when you, when you think about just overall across society, how many people you interact with each day that you feel have high degrees of empathy, and there's going to be a lot of people, you're like, no, that they don't fit that category.
[00:44:36] Paul Roetzer: AI could help add empathy, and I don't know that that would be a bad thing, like Right. Maybe people who aren't empathetic could learn empathy. By being coached by the AI to be more empathetic, like, man, I could see that as a big thing in management, like within companies, if you have managers who just struggle to connect with people.
[00:44:55] Paul Roetzer: I just, I did a talk keynote. And you know how millennials may need more empathy from their managers than baby boomers did. . Or Gen X. And so that kind of thing was where you start again, where you think about the innovation that could occur and the positive impact AI could have. You start to think about, I mean, maybe there is a AI for empathy.
[00:45:20] Paul Roetzer: Tool to be built that lays over all this other stuff and just teaches humans how to be better humans. Like it's, it's doable. I think somebody should build that actually, if you do.
[00:45:37] Mike Kaput: So we also got another major fundraising announcement in AI this week. So Rep, which we've talked about before, is an AI powered software development platform, and they just announced an extension to their previous series B fundraising. And so they are now valued at 1.16 billion. And in this latest extension, they raised a little over $97 million.
[00:46:03] Mike Kaput: And they've specifically stated the funds are to do two big things, expand their cloud services and build their lead in AI for software creation. As part of their platform, they have ghostwriter, which is an AI powered coding co-pilot essentially. So I know you followed Rep for a while. Paul, what was your take on seeing this amount of money being raised?
[00:46:24] Paul Roetzer: Yeah, we've talked about them a couple times before. This company is no joke. I mean, there's, there's some AI companies. I just feel like they're, they're going to be major players in the future and re is one of 'em. . I mean, I met their co-founder and ceo. He's. Brilliant. You know, grinded for 10 years, building this company before anybody cared.
[00:46:47] Paul Roetzer: Moved from Jordan to build it in the us. Like this awesome story of entrepreneurship and perseverance is.
[00:46:54] Paul Roetzer: What he tweeted about it, but they're trying to democratize coding, basically the ability to build apps and companies. And so I just feel like they're going to play a major role in the explosion of innovation and entrepreneurship, and I don't think anybody's aware of that yet. Like in the business world or in the tech world.
[00:47:17] Paul Roetzer: Like they're sort of just coming on a theme and be a bit of a coming out party from, you know, re achieving that unicorn status. The billion dollar plus valuation. Yeah. A lot more people are going to pay attention to what they're doing. But them and like runway, we talk a lot about, we talked about them earlier, like there's just some of these companies that probably fly under the radar for most business leaders.
[00:47:39] Paul Roetzer: Yeah,
[00:47:45] Paul Roetzer: really cool stuff.
[00:47:48] Mike Kaput: All right. We saved a big topic for last year in a very recent development. As of the time of recording this podcast, Geoff Hinton is one of the godfathers of modern artificial intelligence, and for the last decade, he has worked at Google to help them lead in ai. Well, that ended. This week with a bit of a bombshell because Hinton just quit Google and he says he specifically quit because he wants to speak up about.
[00:48:20] Mike Kaput: AI risks and he told our friend Cade Metz at the New York Times that he fears major companies are no longer developing AI responsibly. And he also thinks AI is going to cause some very major issues by things like we just talked about, the generation of fake media and job loss. So Paul, am I right in saying this a pretty big deal?
[00:48:42] Mike Kaput: Yeah,
[00:48:43] Paul Roetzer: it. If you, so, if you haven't read Genius Makers by Cade mats, just go read that. Like you'll understand the significance of Geoff Hinton to, to modern ai, deep learning, everything we're seeing with image amd language generation. A lot of that can be attributed to the work Geoff Hinton over the last 40 plus years.
[00:49:04] Paul Roetzer: So three genius makers by Cade mats. Follow this story. Follow Geoff Hinton. He tweets very rarely. Maybe a couple times a month. I have alerts set up for any time Geoff Hinton tweets, so I'm, I know when he tweets, and that's how I actually saw this, this morning. So just, yeah, I would follow this story and I would listen to what he has to say.
[00:49:30] Paul Roetzer: There's very few people alive who know. More about AI than Geoff Hinton does and what's possible and the near and long-term implications of it. So I would pay attention and I would read Genius Makers by Cade Metz, and we'll put in a bonus show notes. I interviewed Cade Metz for our 2021 Marketing AI conference.
[00:49:51] Paul Roetzer: We did a fireside chat about genius makers, and that video is available on YouTube, so we'll post. That interview and the show notes where we actually get into the story of Geoff Hinton. It's awesome. And the interview with Cade was one the highlights of running a conference for me for the last four years.
[00:50:06] Mike Kaput: Fantastic. Well, you know, I think we've covered just a few major topics today. Paul, as always, thank you for your time and insight. It's incredibly helpful to unpack what's going on in the world of ai.
[00:50:20] Paul Roetzer: Yeah. And, safe travels on the speaking circuit. Maybe we'll see each other sometime in May. Yeah.
[00:50:26] Paul Roetzer: Different parts of the world at all times. But I'll actually, I'll probably run into you in Cleveland. I'll be back for day or two before I next. So I think I might. At your talk on Wednesday. So yeah, I'll, I'll see you soon. And, okay. To our listeners or you know, viewers on YouTube, again, all, all these podcasts are published on YouTube as full videos as well.
[00:50:46] Paul Roetzer: So if you'd rather see our faces in my dimly lit hotel background today, you're welcome to check us out on YouTube. But yeah, we hope to, you know, meet more of you at Macon in July. Like we'd love to see you at the conference and, You know, I've, I've had a lot of chance in recent months to interact with podcast listeners at the events I've been speaking at.
[00:51:05] Paul Roetzer: People come up and say to listen all the time, and it's awesome. So definitely, you know, look forward to hopefully meeting more of you in person. Don't hesitate to reach out to Mike and I on LinkedIn. We're both pretty active there. And, yeah, until next week, we'll talk to you then.
[00:51:23] Paul Roetzer:
[00:51:23] Paul Roetzer: Thanks for listening to the Marketing AI Show. If you like what you heard, you can subscribe on your favorite podcast app, and if you're ready to continue your learning, head over to www.marketingaiinstitute.com. Be sure to subscribe to our weekly newsletter, check out our free monthly webinars, and explore dozens of online courses and professional certifications.
[00:51:45] Paul Roetzer: Until next time, stay curious and explore AI.