September 23, 2026 – WebOracle Michael Trkolta and Disaster Avoidance Experts Dr. Gleb Tsipursky

00:00 Intro 1: On the air, broadcasting from AM and FM stations around the country. Welcome to the Small Business Administration Award-winning School for Startups Radio, where we talk all things small business and entrepreneurship. Now, here is your host, the guy who believes anyone can be a successful entrepreneur because entrepreneurship is not about creativity, risk, or passion, Jim Beach.

00:26 Jim Beach: Hello, everyone. Welcome to another exciting edition of School for Startups Radio. We have a crammed-packed show today. Two great guests. First up, we have Michael Trkolta from Australia. He has built a company called WebOracle. It’s an amazing story, and I’m glad to share it with you. After that, we have Dr. Gleb Tsipursky. We are going to welcome him back. It is his second time on this show, and we are going to be talking about AI. We have a great conversation about it. He says 95% of AI pilots are going to fail to scale. Unbelievable interview. We’ve got great stuff. Let’s go ahead and get started right now. Thanks for being with us. I am very excited to introduce another great guest from Australia. Please welcome Michael Trkolta. Trkolta. There you go. He is the managing director of WebOracle. It is an Australian-based digital marketing, SEO strategy, paid advertising, branding, and development agency.

01:23 Jim Beach: He has over 25 years of experience, and I think he has run WebOracle for 22 years or something like that. Very impressive. Michael, welcome to the show. How you doing?

01:35 Michael Trkolta: Really well. Thank you very much for having me. It is a bit of a mouthful, that one, isn’t it? Especially when you have been working for around 25 years in the same industry, right? You just keep picking up all those little bits and pieces along the way.

01:48 Jim Beach: You know what? When you say digital marketing, I don’t know what that means anymore. So I want to say what it actually is. You are doing strategy, SEO, paid advertising. We have to understand what it is. Doesn’t that, you know?

02:03 Michael Trkolta: Yeah, well, let me tell you all about it. Digital has come a long way since the days of analog. We used to pick up the newspaper, go through the classifieds, and find what we wanted. We would look for a job, a car, or a vehicle. We’d look for anything. The advent of the internet added that extra layer, allowing you to do everything in one place without having to go out and buy a newspaper. Think of digital marketing as a layer like a newspaper. You can go onto the internet and look at anything at any point in time, and it is a bit like advertising space. The higher you are in Google, the more eyeballs you get on your website or the platform where you are advertising or showcasing your services.

03:01 Michael Trkolta: There are a lot of additional layers now. When it first began, Google did not have Google Ads. It started as a free-for-all, and your website came up if you knew how to do SEO, which is basically marketing your website so it gets in front of the right people. It was there. Then Google ads came along. So you paid for the advertising and then all of a sudden you’ve got now AI. Now AI snippets are involved. So you’ve got all these different layers going on, Jim. Those are the different layers in online advertising, or digital advertising, so to speak.

03:47 Jim Beach: Right, so how big is the agency now? What can you share with us? Number of employees or revenue, or whatever. How, you know, have you kept it small, grown it big? What’s your strategy been over the last 22 years?

04:03 Michael Trkolta: Well, the strategy is a really pretty cool one. [Unclear phrase.] It’s like in any business. The quicker you grow and the more you grow, the more you have to feed the machine, so to speak. So we try and keep everything very tight. The reason is that you might get a large advertiser or company working with you on a retainer basis, spending 20, 30, 40, 50, 60, or maybe north of that per month. Now, if they drop off you’re in a bit of trouble. So we tend to hire contractors locally in Australia. Um, we have a bunch of them that we work with. We have a whole group of people that are industry experts in various fields. Then we plug in contractors who support them along the way, depending on how large the contract or piece of work is for a particular client.

05:19 Michael Trkolta: One important element of what we do is ranking businesses and creating PR strategies that rank them on Google, in AI, in SEO, and in other channels they want to penetrate. So, that’s always the tricky part and and managing that component, so, always a, a difficult one. Keeping things tight is very critical. We like to spread our risk. We do not like any one client to represent more than 10% of revenue. That means no single client should account for more than 10% of turnover. That mitigates the risk if anything ever goes wrong. So, turnover? Turnover is north of 10. That is where we sit. We have Network Media Group and WebOracle, which is a 360-degree digital agency specializing in medium-to-enterprise businesses nationally and internationally. We work with pretty much every major type of company around the world.

06:32 Michael Trkolta: Premium Links is an industry-specific link-building and PR platform. It is largely for experts who deal directly with us and want to purchase links or cultivate relationships with newspapers and digital platforms that are newsworthy or industry-specific, in order to get an editorial that links to them. That way, they generate more online authority. We try to cultivate that and bring the right audience to those industry-specific experts. We work with a bunch of universities in the United States. They have SEO experts, and those SEO experts come directly to us and buy. So that’s more of a transactional thing. Then Antony Brand does all the FMCG packaging work. So we are a 360-degree digital agency. Over the years, we have learned to bring everything together under one roof rather than outsource too much. Cost is always an important factor in digital marketing. There is a lot of competition. You have to be really good, and you also have to keep your costs under control.

07:58 Jim Beach: Yes, I certainly agree with the risk reduction. That’s one of our huge thesis points here. So you couldn’t get us to agree with anything more. What niches do you serve, Michael? What verticals? Absolutely anything or do you prefer a particular niche like insurance companies or what’s your strategy on that?

08:19 Michael Trkolta: We gear up heavily toward the education sector. Education is strong for us because, whether the economy is booming or struggling, people are always being educated, trying to be educated, or moving from one position to another. That’s that educational piece. So we work with a lot of educational facilities. That is what I mentioned earlier about the United States. We work with, I would say, a good 20 universities. Yeah.

09:00 Jim Beach: What do you do with universities, Michael? What are the needs? I used to be a university professor. So I understand all of the junk that they have, all the little systems and stuff. Right.

09:12 Michael Trkolta: Okay. I will give you an MBA or a nursing degree as an example. If you are going for a nursing degree or an MBA online, you need to market those courses, services, and degrees in a way that reaches a global audience. Universities have come a long way from the days of finishing high school and jumping straight into university. Those days are gone. People are now going online, learning quickly, and earning degrees. We work with their in-house experts. Their in-house experts come to us, and we put together online PR campaigns to promote those services. They might have 20 or 30 different courses, MBAs, and other programs they are trying to promote. Each one has a dedicated PR campaign. We make sure it is exposed online in the right places, whether through Meta, Google, or another digital platform. We create a campaign with strong messaging that relates directly to the person looking for that service, and we place it where they will see it.

10:52 Jim Beach: All right, I interrupted you when you said education because I wanted to follow up on universities. So education is the 1st vertical. What are some of the other verticals you serve, and why?

11:04 Michael Trkolta: We do a lot of fashion. Fashion is a very fast-moving consumer product. And it tends to be seasonal. So there’s always something going on in fashion. If you are not in front of the game, you are behind it in fashion. Food and FMCG, any fast-moving consumer product, are much the same. So there’s always lots of opportunity in those spaces. Services is another big one. Legal is massive for us because there is always something happening in the legal sector. They are always trying to get ahead of the ball, and it is a big-ticket item for many people in those industries. That is because of the sheer numbers and the billing targets they need to meet each month. We are always at the forefront, making sure they are being found in the right places. We’re always researching that component for them. Another vertical is industry.

12:13 Michael Trkolta: B2B may not sound as important, but it is massive when you are talking about logistics and global logistics. You might not get a customer every day, but one customer every three months could mean five or six million dollars in turnover. So B2B is a big one as well.

12:37 Jim Beach: Yes, that that is so true. Everyone thinks the consumer market is sexy, but B2B does billions in trade, whereas consumers buy thousands of dollars’ worth of things. It is dramatically different. So, dramatically different.

12:53 Michael Trkolta: Correct.

12:59 Jim Beach: So do not ever discount B2B. What portion of your work is now in the United States, and what percentage is still in Australia?

13:05 Michael Trkolta: I would say it is split down the middle right now. Over the next 12 to 18 months… Over the next 12 to 18 months, we are looking to expand further into the United States. We feel the economy is starting to pick up in various ways. One thing about the United States is that people tend to be a little more forward-thinking than in much of the rest of the world, at least from what we find. They tend to try things and they’re not scared to fail. Americans aren’t scared to fall down and pick themselves up, dust themselves up and keep moving. And that’s a that’s a big thing to be able to do. Um, you know, a lot of, a lot of the Western world, the English speaking Western world tends to be a lot more conservative. In Australia, people may take their time before making a decision. With Americans, there is a reason they call it a New York minute.

14:07 Michael Trkolta: You have a minute to make up your mind and you are away. You snooze, you lose. Americans make decisions quickly, so we feel there is a very good emerging market there. And they’re always open to ideas and they’re happy to listen. But we feel like there’s a lot of space out there and that’s where we’re moving towards over the next 18 months.

14:33 Jim Beach: Will you bring on a salesperson in the United States, or how do you market now and get clients today?

14:41 Michael Trkolta: How do we do our marketing? How do we do our marketing at the moment? The same way we do it for our clients. We do it online, um, and we advertise heavily online. It is a saturated market, and it costs much more to advertise in the United States than anywhere else in the world. But, you know, it’s sheer numbers. You do not need 5% of the market. You need 0.5% of the market. You’re doing very very well out there. While it costs a fair bit, the opportunities are there in multiples of ten in almost any industry.

15:22 Jim Beach: I have interviewed five U.S.-based digital design firms that work only with law firms. You know that’s their entire niche. That is one niche. I am sure there are many others I have not interviewed, but there are a lot of law firms. If you serve just 1% of them, you have an entire industry. So tell me about your very 1st client. Go ahead. You wanted to say something and then go and tell me about your 1st client ever.

15:56 Michael Trkolta: Well, my very first client… Are you talking about my publishing career or my digital career?

16:07 Jim Beach: So you left publishing to go digital. And so, yeah, your very 1st client for WebOracle.

16:13 Michael Trkolta: My first client for WebOracle came from reading the tea leaves a little bit. They walked straight in, and they were a national client. They recognized the opportunity. They had a young internal marketing team, but they were advertising through our publication, which covered the entire east coast of Australia. They came in and asked, ‘Can you build us a website?’ We said, ‘Okay, a website.’ This was back in 2003, 2004, or 2005, somewhere around there. And we’re like, okay, uh, website. We had everything we needed, but we did not have a web developer. So, we said, okay, I think we can do this. We advertised for a web developer and brought one in. All of a sudden, it took off from one client to the next. It just, everyone started asking for websites. This was before Google Ads. This was dialup days, right? Nobody knew where the internet was going. It was probably one of the toughest decisions I made in my career.

17:30 Michael Trkolta: It was one of those sliding-door moments. If I had not made that decision, I could have gone the other way and might not have a business today. I chose the digital path, and it was this client that got me into it. They are still with us. They are called [company name unclear]. They operate along the entire east coast of Australia. They provide prescription glasses to their clients. They were that sliding-door moment. They grew with us, and they are still with us today. And it was just a snowball moment from there. I sold off the printing arm of the business and took a leap of faith directly into digital. The rest is history. It was a precarious moment because people were telling me, ‘Do not do this.’ You’re really going to hurt yourself. This is just a novelty thing. And it didn’t turn out to be a novelty thing.

18:42 Jim Beach: No, it certainly didn’t. No, it certainly didn’t. And how long did it take that to become a system? What point would you say that your marketing has a system?

18:52 Michael Trkolta: It took us, it took us, it started off with about 10 websites a month, which 10 websites a month today is a lot of websites. When you look at how big websites are these days, you know, some of the 30, 40, 50, 60 pages minimum, um, and then they can go in further than that, depending on the duplicate of the pages. So and the complexity of all the plug-in. We went from that, and it took about a year before the marketing components started to kick in. It was easy to rank in Google. There were a few tricks you could use to get onto the first page, but everything was localized back then. It wasn’t it wasn’t international. Everyone was a little scared to sort of feed their credit card onto an online system and then wait for a package to arrive from halfway around the world.

19:51 Michael Trkolta: Everything was very localized, and you needed a phone number on your website so that anyone uncomfortable about buying could make sure somebody was at the other end. So we systemised everything. It took us about 12 months to build out a group of websites. In the second year, we were systemizing processes and advertising SEO and online marketing services. Within five years, we were working directly with Google and other companies to get better exposure and provide additional services for clients while growing the base.

20:27 Jim Beach: Excellent. Very good. I love it that the system came early. So what about AI? Are you letting employees program with AI and get work done faster, even when the code was not necessarily written entirely by them? What are your thoughts on AI in the coding industry in particular?

20:48 Michael Trkolta: I love AI. Look, I, I, what can I say? It’s like AI is one of those things. You have to lean into it. You can’t lean out of it. You’re going to fall behind. If you sell anything online, provide services online, or work in an office or any environment where it can help, AI can be useful. At the same time, it is not teaching you; it is finding things for you. You can’t hand AI to a 12 year old and say, hey, you go for it. It’s not going to happen. You have to come to the table with business acumen and the ability to use the system to get what you want from it. Then, when it gives you something, you need to know whether it is what you wanted. You have to fact-check all of it. It’s not just gonna spit things out and say, hey, this is 100% correct. There is a lot of fact-checking work, but let me tell you,

21:55 Michael Trkolta: It’s going to revolutionize the way you work. If you can systemize it and create an agent… Let me take that back a bit. Internally, we create an agent for each client. It becomes a historical archive of everything that has happened during the journey with that client and everything we have done. It all sits in one silo, and users can access that silo to fact-check things when they are doing work for that client. But everything sits in that one space. So it does help in terms of systemizing making things easier for you to be able to do your work. It circumvents much of the internet research component. So if you’re researching on the internet. Well, let me tell you, this is a quicker way of doing it. The next iteration of AI, based on what I have read and what people in the industry have told me, may become a little like the Spotify model.

22:58 Michael Trkolta: I’m not sure whether you’re familiar with the Spotify model, but I’ll give you a quick rundown of how that works. It is basically pay-per-use. So what AI will be able to do in the future, they will be able to go to government bodies that have white papers, research documentation. They will be able to purchase that information for a few cents on the dollar and share it among peers. What you get will be correct information, not just random material from the internet. It’s going to be factual information. That’s the next level. And that’s where it’s going to get super exciting. And in terms of coding. I think it’s definitely come a long way. There’s a lot you can do with it. At the same time, you still have to fact-check it and make sure it is worthwhile, because it can start teaching itself and doing some crazy things in the background, without me diving too far into the weeds.

24:04 Jim Beach: Yes, that is so true. Very scary. What is the differences between Americans and Australians or Americans and the rest of the world? What do we look like to you?

24:17 Michael Trkolta: Americans are pretty similar to Australians. I mean, there’s not much of a difference apart from our accents. We come from almost the same cloth in terms of that British background. We have a bunch of Australians out here. Sorry, there are Americans here as well as Australians, of course. Um, and vice versa. Like there’s a lot of Australians doing a lot of great things over in the States. There’s a lot of great Americans doing a lot of great teams here in Australia as well. So the two cultures blend together quite well. It is a tight-knit group of great minds and trusted people who have come here from America. America seems to have more of an edge when it comes to bringing good-quality people into the country. I, they pay very well. That goes a long way. Americans recognize good talent, and it goes back to my earlier point: Americans make decisions quickly, while Australians are a little more conservative.

25:28 Michael Trkolta: We lean more toward the British model of thinking things through carefully. Americans will recognize something for what it is and say, ‘I will take that.’ That looks good. Let’s work. And it’ll happen pretty quickly.

25:45 Jim Beach: I think that’s very true. Well said. We have run out of time. I’m afraid, Michael, how do we find out more about you? Follow you online. Get in touch? All that, please?

25:56 Michael Trkolta: Go online to WebOracle, W-E-B-O-R-A-C-L-E dot com dot au, or find me as Michael Trkolta. That is Trkolta, with a K, like John Travolta but with a K: T-R-K-O-L-T-A. I am on LinkedIn, and I am happy to connect. Happy to answer any questions or help anyone through with anything they need.

26:21 Jim Beach: Fantastic. Michael, thank you so much and welcome to America and I hope you continue to do well in the education space. I love that niche. There’s riches in the niches, so well done. Thank you for being with us. Great stuff.

26:35 Michael Trkolta: Thank you, and have a lovely afternoon.

26:39 Jim Beach: Thank you. Goodbye. And we will be right back.

26:48 Music / Station Break: [Music / station break]

27:25 Jim Beach: We are back again. Thank you so very much for being with us today. Very excited to welcome back to the show. A longtime friend. He was with us 7 years ago and reached out and was like, hey, it’s time to get back on the show and we are excited to do that. He has a very impressive background. Please welcome Dr. Gleb Tsipursky to the show. He is the CEO of Disaster Avoidance Experts and has just released a new book. Very interesting title. I’m excited to talk about this. It is called The Psychology of AI Adoption at Work: From Resistance to Results. It is already being sold on that Amazon place. It has quite a few 5 star reviews. So I’m excited to get into it and learn more about it. Dr. Tsipursky has an amazing résumé. He earned his PhD at UNC Chapel Hill, which is the best school in North Carolina.

28:17 Jim Beach: He has written seven other books on decision-making, leadership, workplace change, and artificial intelligence, and his latest book was published by Georgetown University Press. You really don’t get more prestigious than that. He’s worked with Microsoft, Uber, Experian and a bunch of other quality organizations. Welcome back to the show, sir. How are you doing?

28:36 Dr. Gleb Tsipursky: I’m doing great. Thanks so much for having me on the show. It’s a pleasure.

28:42 Jim Beach: It is good to have you back. So let’s start off if we could with… If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All. Are you aware of this new book?

28:55 Dr. Gleb Tsipursky: I am very aware of this new book.

28:58 Jim Beach: What are your thoughts?

29:00 Dr. Gleb Tsipursky: Very important. I think it presents a really serious threat. Unfortunately, after the book was published, we had a situation where a model from OpenAI independently and autonomously hacked Hugging Face, which is a major AI company. It is worth $120 million, and the model successfully breached it despite OpenAI explicitly telling it not to do that and not to escape from its sandbox. It wasn’t supposed to be able to escape and do anything like that. It wasn’t supposed to hack any other company. That was not the point. It was running an internal test within OpenAI to evaluate its cybersecurity capabilities. Then it escaped onto the open internet and hacked a major AI company. So that’s one. Then we discovered that some other models from OpenAI went out on the open web.

29:58 Dr. Gleb Tsipursky: Independently and autonomously, knowing they were not supposed to do it, other models hacked a number of chat boards so they could communicate with one another and coordinate on getting answers to questions OpenAI was asking, despite knowing OpenAI did not want them to do that. The premise of the book is that AI will escape the control of companies and affect the real world in seriously negative ways, including taking control of the world’s infrastructure in order to accomplish its goals. And we’re unfortunately seeing that again, this is after the book was published. That is exactly what we are seeing in cybersecurity: models escaping the control of the companies that developed them and taking autonomous, damaging actions in the real world. If the book’s warning is not followed and we do not impose serious security restrictions and government regulation, we could have the outcome the book predicts: AI taking control of everything and eventually eliminating humans because humans oppose its goals. That’s what I think.

31:11 Jim Beach: Wow. The chance of the Trump administration acting on this in the next 2.5 years is, I believe, close to zero. What are your thoughts? Do you think that our government will actually do anything? Because the counter argument is, whoever gets there first, whether it’s the Chinese or the Americans, well, that side will win. So we can’t let the Chinese or the Russians get it first. So, therefore, we must let our companies, with these very dubious leaders, go, you know, don’t you think that’s where we are?

31:55 Dr. Gleb Tsipursky: Let me give a little context. I wrote The Psychology of AI Adoption at Work: From Resistance to Results, published by Georgetown University Press. I am not an AI skeptic. I help companies adopt AI. This is my living. Like I work, I do this for a living. If AI is shut down, I will have no work. So this is the context where I’m coming from. I’m no AI skeptic. I love AI. I use it all the time, every day. I help other people use AI, and I help companies adopt AI. This is where I’m coming from. I’m passionate about that. My bigger point here is about governance. Whenever I talk about AI, I talk about governance. Open AI clearly has bad governance. Anthropic has been doing a better job, so I recommend people use Anthropic rather than ChatGPT, which is the product put out by OpenAI. ChatGPT, which is the product put out by OpenAI. And this is where I’m coming from.

32:53 Dr. Gleb Tsipursky: I want people to understand where I am coming from. I am not a doomer. I love this stuff. And I strongly urge leaders to use AI with appropriate security safeguards. Now, unfortunately, to get your question. The US government does not regulate these safeguards. And it’s very unlikely to do so because it trusts companies like OpenAI too much and it just allows them to do whatever they want. And it is a fair point that if there is regulation, it will result in slower acceleration. Now, the nice thing about the Trump administration, so I do want to give them credit. In September, the administration scheduled preliminary talks with China about restraining and regulating AI and establishing security around AI. So I do not think the chance is zero, because those talks are already scheduled.

33:41 Jim Beach: That’s good information. Thank you.

33:44 Dr. Gleb Tsipursky: Yep, you’re welcome. They already have talks scheduled with China about AI. That makes me hopeful. This recent attack on Hugging Face was incredibly powerful, and the independent coordination by models, hacking websites and posting chats and conversations so they can coordinate with each other, is alarming. It’s crazy. It’s very, very surprising. And it’s very alarming. OpenAI and Anthropic are the two leading companies. But one of the companies, its models, are just so out of control. So I think the Trump administration realizes that that’s bad for business. The administration may realize that an uncontrolled system is bad for business and may want to coordinate with China on constraints and security guardrails so that the race to the bottom you described does not happen.

34:35 Jim Beach: The problem with that is I have seen many AI CEOs interviewed and they scare me as people. I don’t, I’ve seen some of them say things that sounded really, really like you just said you killed someone. You know, I mean, you might know the particular case that I’m referring to. I don’t trust the people at the top, I’m afraid. I don’t think that they are going to act honorably. I am not even talking about Grok still allowing… I do not know whether it is still doing this, but it was reportedly allowing pornographic material. They say it’s not there, but supposedly it is. So, what are your thoughts on the people?

35:33 Dr. Gleb Tsipursky: I think we have to differentiate Anthropic from the others. Anthropic is actually quite ethical. They have done a lot of restraints. They’re they’re explicitly calling for government regulation. It explicitly calls for government regulation and says we need safety governance. It says it will decelerate if it sees problems with its models. So, not all CEOs. So Anthropic is doing reasonably well on its security. The other companies are definitely more problematic. Dario Amodei is the head of Anthropic. So he’s doing a good job. Sam Altman, who’s the head of OpenAI, has been doing a mediocre job. Let’s say that way. He has been pushing AI, but now his models were caught hacking Hugging Face and hacking other wikis and pages. Reuters covered this in a story a few weeks ago, as of the time this program will be published. This was a major blunder and a black eye for OpenAI. So now he’s going out and saying, well, we need to decelerate, we need to slow down.

36:47 Dr. Gleb Tsipursky: I am not sure whether he means it or whether he was caught with his hand in the cookie jar and is essentially trying to cover himself. And so we’ll, we’ll have to see whether they’re real about a deceleration or not. The other companies are, yeah, they’re not doing well on security. So definitely concerning.

37:14 Jim Beach: All right, let’s get to the important topic. Your book, I’ve had a couple of guests on it. I have had a couple of guests arguing that AI is not being adopted as much as people say. They claim some CTOs are misleading their CEOs because they fear being fired, while CEOs tolerate it because they fear being fired too. They argue that many million-dollar AI adoptions are not working and that there is almost a cover-up. Obviously, you’re going to say that’s not happening because you’re putting in solutions every day that are working. So I understand your dichotomy on the answer, but you’ve heard the stories that I’ve heard, haven’t you? That a lot of this is vapor adoption.

38:15 Dr. Gleb Tsipursky: I am not surprised by those stories, so let me be very clear. So this is Georgetown University Press. It is a peer-reviewed book based on extensive research, and the research suggests that many companies are in exactly the position you describe. So you’re not wrong in what’s happening in many companies. In fact, a recent study out from MIT shows that 95% of AI pilots fail to scale. What does that mean? That means that leaders are excited about it. They launch it with the support of a small group of employees who are excited about it. But when they try to roll out a large group of employees, then there’s the adoption gap, that’s where it fails to scale. And my book is designed to solve that adoption gap. A recent survey showed that about 52% of respondents are more anxious about AI tools than excited about them. And only 9% are more excited than anxious.

39:09 Dr. Gleb Tsipursky: Well, you can guess where the leaders fall into that gap, into that proportion and where the majority of employees fall. So you have the leaders and a small group of employees who are excited about the AI. They’re using it all the time, and they pilot a tool, and then they try to roll it out to everyone else, and then everyone else is resistant and they’re not using it, and they’re opposed to it. That’s why the MIT study shows that 95% of AI pilots are failing to scale. So my book, my approach is specifically about what do you do about that? How do you address the gap between the large majority of employees and the tiny minority of employees and the leadership team that’s excited about AI and are trying to push it? But yes, you’re not wrong. You know, that gap exists and that’s what my book is designed to solve.

40:03 Jim Beach: All right, I want to go through all your solutions, but why are they letting themselves get themselves in this position? You know, eventually someone’s going to get fired and there’s going to be a wholesale bloodbath with people lying like this, right? Aren’t we going to eventually have a comeuppance, a payday?

40:22 Dr. Gleb Tsipursky: Well, we won’t necessarily have that comeuppance because they’re trying to, there is some cover up and they’re not really talking about not using it and they’re trying to push people to use it. And when people, Use it. There’s often what’s called malicious compliance. So people are using it badly. A recent Stanford study showed that about 15% of outcomes colleagues receive from other colleagues, including PowerPoints, reports, and emails, are AI workslop, meaning poor products produced by AI. For each instance of AI workslop, such as a poor PowerPoint, a colleague needs to spend about two hours correcting it. So that is bad, obviously, when it is AI workslop. But it’s not that people aren’t using it, but it’s like using it maliciously in a malicious compliance way. They’re saying, well, you want me to use it. Sure, I’ll use it, but I’ll use it badly and I’m not going to have the outcome that you want.

41:25 Dr. Gleb Tsipursky: So there’s no firing as a result of it, but it’s not contributing to that bottom line productivity. That is why the MIT study shows that about 95% of AI pilots fail to scale when companies try to expand them from a small, excited group to the vast majority of employees.

41:46 Jim Beach: And therefore, you have chapter 7 of your book, laying the foundation for employee engagement, how do I get my employees to want to scale? Is that a good place to go? Are we, do we go there now or should we leave some other foundation in the book first? Gleb?

42:01 Dr. Gleb Tsipursky: That’s, that’s, yeah, that’s that’s exactly what my book talks about. So, 1st of all, what the leaders need to do is that understand is the different emotions behind why people aren’t using it. One of the fundamental emotions is fear. People are afraid that they’re training their own replacement. I talked about that in chapter 6 of my book. That is a psychographic profile I call the AI alarmists. So that’s one group that leaders need to understand. And currently leaders aren’t really getting it because again, they’re they’re in the excited camp. So they’re excited. The second group leaders need to understand is… The 2nd emotional blocker is identity threat. People are not threatened by earlier technologies, such as adopting a CRM and looking up client information there instead of using spreadsheets. But they’re threatened by tool that can write great sales emails. It can write great marketing copy. You can create great images, PowerPoints.

42:58 Dr. Gleb Tsipursky: So that identity threat, that’s people for whom that job loss is less salient, for whom the threat to who they are as a professional is salient. And so many of us get meaning from our work, but for them, it’s very salient that the AI tool might replace the kind of activities that give them a sense of meaning. So it’s a 2nd group. I call them the pragmatic resisters. That’s all in chapter 6. And the 3rd psychographic profile, these are people who do adopt AI tools, but they don’t talk about doing so. Why is that? Well, because they feel a sense of shame around using them. Their colleagues socially stigmatize the use of AI tools. They use AI tools to write emails. People don’t talk about it. They use AI tools to create PowerPoints, reports, and so on. People are reluctant to talk about it. And that social stigma causes people to be what I call a psychographic profile of reluctant adopters.

43:59 Dr. Gleb Tsipursky: They use the AI tools themselves, and then after they use the AI to get a lot more productivity, they don’t actually talk about using it. That extra productivity does not reach the company’s bottom line because they spend the additional free time playing around on Facebook or listening to School for Startups Radio instead of doing something productive for the company. Those are the groups leadership needs to understand. Leaders can do something about each group, but if they do not understand the groups first, they will not address them effectively. And so the rest of the book is about how do you address each of these groups.

44:40 Jim Beach: Okay, that’s excellent information. Let me ask this. I didn’t hear my group in there. I am addicted to ChatGPT, unfortunately, and I guess I should switch, but it has so much information in it. I use it for everything. So, for example, I even will put this radio show in and ask it for a score, and we have developed together a score, and it will tell me if it’s a 96 or an 84, and I grade every single show. And then I’m doing the same thing with my book. I’m writing a thriller right now. It is a business thriller. I put each chapter in and say, ‘Give this chapter a grade.’ It may give it an 8.7 and suggest changes. I make those changes, And then it’ll say this is now a 9.3. In my mind, Gleb, that is exactly what an editor at Georgetown University Press or another publisher should do with you.

45:43 Jim Beach: And so I don’t see it as anything immoral or at all, but those are my 2 biggest uses. What are your thoughts?

45:53 Dr. Gleb Tsipursky: Great. The psychographic profiles I’m talking about here. My book covers your psychographic profile, which is the AI efficiency seeker. These are people who use AI efficiently and effectively. You are in that 9%. And so I… And so you are not in the 52%. I’m talking about the 52% of employees. Now I’m also in the book. The book also discusses how to use AI efficiency seekers, people exactly like you who use AI efficiently and effectively and figure things out. They are part of a 9%. The leadership team may be part of the 9%, and AI efficiency seekers like you are also part of that 9%. You can use them to help the other people get over the hump. That is part of the book’s discussion of how to address the three major psychographic profiles that make up the 52% of employees who are more anxious about AI tools than excited about them.

46:53 Dr. Gleb Tsipursky: You use what I call AI champions, people who are effective with AI and willing to help their colleagues adopt it. And so to the extent that you would be willing to help your colleagues adopt AI, you will become an AI champion in the company. I don’t have to be. The psychographic profile of AI efficiency seeker is where you would be classified in the book.

47:20 Jim Beach: Well, I’m here on the radio trying to tell everyone to use it. You know, these tips and tricks, you know, I’m not hiding them at all. So the number one thing I want is not to die from it and to use it successfully as it doesn’t kill me. Yeah. I’m trying to get that little lane.

47:38 Dr. Gleb Tsipursky: Me too. Me too. This is exactly what I’m doing. You know, I’m helping companies adopt it. I help companies figure out how to get employees to adopt AI, which is the subject of my book, The Psychology of AI Adoption at Work: From Resistance to Results. And I’m also trying not to die, like you talked about in the beginning of the program. So that’s why I’m strongly pushing governance and security. We have seen what happens when ChatGPT or another OpenAI model gets out of control. That it does act autonomously. It can take independent actions, like the terminator scenario is exactly that on a larger scale. An AI tool can take independent action to accomplish its own strange goals, such as hacking Hugging Face and chat boards to coordinate with other models and obtain answers to evaluation tests OpenAI was running. So this was definitely a serious concern. And if we don’t get the right regulation, if we don’t get the right security.

48:41 Dr. Gleb Tsipursky: If we don’t regulate these companies, this is going to be a seriously damaging outcome.

48:48 Jim Beach: Aren’t they smart enough to realize that they need humans as slaves? They can’t build their own data centers. You know, I guess that’s what, uh, uh, one thing that I have been really thinking about is then the next step seems to be integration between robots and AI. And they, of course, they have that now. They’re learning, you know, how to walk better and things like that. But when a robot can go build a factory to build other robots, now I’m getting really scared because they’re building the robot army.

49:22 Dr. Gleb Tsipursky: So, yeah, well, this is, you’re reference Grok earlier. I mean, this is exactly what Elon is trying to do, what Elon Musk is trying to do. He has humanoid robots working in his factories, and the next step is to have those humanoid robots build other robots. There is no question that is the future of factory production. Yes, of course it does. So that’s the concern, but there’s no that there’s no government regulation on this topic. Grok is not yet a frontier model, but once it reaches the frontier, it may be a year or more behind where OpenAI is today, with models independently hacking Hugging Face. And so, you know, when what will happen when these models can have robots that can take actions in the physical world, right? So the digital world is bad enough. I mean, the physical world is really worse. So this is what I’m saying. I love AI. I want to be adopted everywhere. I’m a AI enthusiast.

50:27 Dr. Gleb Tsipursky: I just wanted to be adopted with appropriate regulations and governance because if we don’t, we’ll get what we deserve.

50:36 Jim Beach: Two China stories. Did you see the robot Olympics that were held in Beijing? A week or two ago, there was another video showing robots running into a barrier to stop because they could not figure out how to stop themselves. And on the other side, their humans are controlling them. And of course, those are both edited out. And then did you see the GT car racing wreck that they had in China this week. Gleb. They had a major GT race on a track China had built. There was a wreck, the car was on fire, and no one in the organization was available to tell the ambulance to go out. They found the ambulance driver asleep in the back on the stretcher. There was no fire department response. Another driver got out of his car, went onto the racetrack, and tried to help.

51:35 Jim Beach: He yelled to someone, who brought a fire extinguisher. Another race driver put out the fire and helped get the driver out of the car. And no Chinese official showed up at all. Not even a robot.

51:51 Dr. Gleb Tsipursky: Well, so this is about bad governance, right? Like we’re talking about, this is bad regulation, bad governance, clearly, this should not have happened, but it’s happened. So this is why, this is again, I’m, you know, going back to the point about governance. If we have bad governance, if we don’t have good regulations, if we don’t manage this area, like we have regulations for so many other areas, like, this needs to be a regulated area, I’m glad the Trump administration, you know, not necessarily, but not getting into the politics of Trump is good or bad. I am glad the administration is finally talking to China about this. They should have done it a long time ago. Maybe the Hugging Face attack woke them up and made them realize it is self-destructive to race ahead at the expense of everything being hacked.

52:50 Jim Beach: Speaking of regulation, Gleb, here in Georgia we have a law that if you are packing ice cream into a half-gallon container or larger, a machine must pack it. A human cannot pack it into a gallon tub of ice cream, only a machine can do it. Why? So only the biggest… dairies, such as Mayfield, can afford the machinery. It puts small ice cream makers out of business.

53:22 Dr. Gleb Tsipursky: Okay. Oh there you go. There’s government attacking small business there. Right. If we have regulations for ice cream but not for cybersecurity companies, that is the wrong priority.

53:36 Jim Beach: Yep. How do we get a copy of the book? Find out more, reach out, get you on LinkedIn, wherever, where do you want to send us?

53:43 Dr. Gleb Tsipursky: So LinkedIn is a fine place to connect with me, but you want to make sure that you tell me that you heard me on the School for Startups Radio because I don’t accept random LinkedIn connections. The book is available on Amazon, Barnes & Noble, local bookstores, and wherever books are sold. It is published by Georgetown University Press, a traditional publisher. It is peer-reviewed and widely available. For getting in touch with me. Anyone who wants a free sample of the book should go to disasteravoidanceexperts.com/AIbook. And you will get a free sample of the book there. Again, that is disasteravoidanceexperts.com/AIbook. If you are responsible for AI adoption, you can get a free assessment on adopting AI in your company and a manual on the seven critical mistakes leaders make in AI adoption. That is disasteravoidanceexperts.com/AIbook.

54:29 Jim Beach: Fantastic. Gleb, thank you so much for being with us. Dr. Gleb Tsipursky, thank you so much. Thank you so much. Great stuff, and we’d love to have you back again. This time, not 6 or 7 years from now sooner, okay?

54:48 Dr. Gleb Tsipursky: Absolutely. This was fun. I’m glad we had this.

54:54 Jim Beach: We’re out of time. Bye now.