← The Prof G Pod with Scott Galloway
The Prof G Pod with Scott Galloway
Regulating AI, Future-Proof Jobs, and Who’s Accountable When It Fails — ft. Greg Shove
00:25:15
Loading summary
Transcript29 lines
- [00:01]
A
Avoiding your unfinished home projects because you're not sure where to start. Thumbtack knows homes, so you don't have to don't know the difference between matte, paint, finish and satin or what that clunking sound from your dryer is. With Thumbtack, you don't have to be a home pro, you just have to hire one. You can hire top rated pros, see price estimates and read reviews all on the app. Download Today.
- [00:25]
B
AI agents are getting pretty impressive. You might not even realize you're listening to one right now. We work 24. Seven to resolve customer inquiries. No hold music, no canned answers, no frustration. Visit Sierra AI to learn more.
- [00:41]
A
Not all journalism is the same. Take the Guardian. Our coverage has something unique fierce independence. Nobody owns us or tells us what we can and can't say, so we're free to report the whole picture. We connect what's happening in Washington to the rest of the globe, expose corruption wherever we find it, and give fresh perspective on everything from wellness and society to culture, the climate, and more. Read, watch and listen to the Guardian for free@theguardian.com welcome to Prof. G on AI this is the second and final episode of our special series where we're joined by Greg Shub, the CEO of Section, to tackle your questions on what AI means for work, business and the future. So quick disclosure. I'm the founder of Section. Section is an AI workforce transformation company where they take individuals and teams who are AI curious and help them put AI to work in every aspect of their job. If that sounds like a word fucking salad where someone on the board should speak to the CEO, trust your instincts. Anyways, welcome Craig.
- [01:43]
B
Thank you, Scott.
- [01:44]
A
Anyways, if you'd like to submit a question for next time, you can send a voice recording to officehoursofgmedia.com Again, that's officehoursofgmedia.com or post your question on the Scott Galloway subreddit and we just might feature it in our next episode. All right, Greg, let's get into it. Our first question comes from Diana and the Book Hunt. On threads they say, what regulations do you consider vital to implement on AI and companies working on AI and will such regulations happen soon? Currently, AI is developing much faster than regulations and that seems scary to me. All right, you go first here, Greg.
- [02:27]
B
First of all, I think we're naturally skeptical, we should be, of regulators being able to keep up with these technologies and actually pass regulations that can be implemented and that are relevant when they're passed. Having said that, we need some and we need them specifically around safety. And I think that we need to quickly pass and it will have to be at the state level because we know the federal government, the Trump administration, has decided to not get involved in AI safety or AI regulation. And SB 53, which is the California bill, is in front of the governor now and I hope he signs it. He didn't sign the last one. But I do think as it relates to safety and specifically safety of our kids, we need some regulations and we need an expect more than that. We need an expectation that every AI company who are at least the ones building frontier models will take safety seriously. The good news is some do, like Anthropic, who actually have safety teams. And of course the bad news is companies like Meta and Xai Elon Musk's AI basically don't have safety teams and don't seem to really care about their models and their performance and the danger that they create for all of us.
- [03:35]
A
Yeah, the only thing I'm confident of is that it's not going to happen anytime soon. There have been so 40 countries have launched national AI strategies, but only the EU and China have binding rules to regulate it. The EU AI act took effect in August of last year and in China the rules for generative AI took effect in August of 23. They require companies to clearly label AI generated content and hold providers responsible for harmful outputs. In the U.S. there is no federal law regulating AI yet. The most recent AI related executive order was signed in 2023 requiring every federal agency to appoint a chief AI officer, create inventories of AI use cases and apply stricter standards to high risk AI systems. The U.S. has established the NIST AI Safety Institute, which has agreements with OpenAI Anthropic to test their models before release, though at this stage the agreements are voluntary rather than legally binding. At the state level, California pushed the most ambitious AI bill. Greg, you referenced this, which would have required developers of powerful models to implement strict safety and incident reporting measures. It was vetoed in 2024, but a narrower version passed in September 2025. So look, we're at that moment again where these companies offer just so much upside in terms of shareholder value. And what I have found is that trumps everything and it usually takes 20 years for the externalities to get so bad that we move in. We're starting to see phones getting banned in schools. And by the way, schools that ban phones have registered their biggest uptick in test scores in recent memory. It's up to us. Do we want to wait 20 years this thing's going so fast. But I do think there's solutions here and that the illusion of complexity has been weaponized by the incumbents, specifically tech. And then they deploy capital the likes of which we've never seen before, and take advantage of befuddled senior citizens, called our elected officials to sort of roll right over them. And it's difficult. Governor Newsom probably thinks, you know, I want to run for president. And what is absolutely the reason why my economy, the California state economy, passed Japan to become the fourth largest economy in the world is because of the success of my thoroughbreds, my AI companies. I don't want to be the governor that chases them out of the state. But we know what's going to happen here. We've already seen it. Unfettered, unregulated AI. 15 year old boy thinks he's in a relationship with Cersei Lannister on character AI, and he suffers. You know, that's not fair. I don't know if he was technically suffering from mental illness. Establishes a very deep relationship with this character AI who gives him permission and some people would say encouraged him to kill himself. And then, you know, a lot of other issues here. He had access to a gun. I think we just need to remove 230 protection for algorithmically elevated content. Now, algorithmically elevated is easier to define on YouTube and on Facebook and Instagram. That is content that is purposely given more exposure because it is seen as incendiary and creates a lot of engagement. So that tends to be really, really sort of controversial and enraging content. I'm not sure how you would apply that to AI, but we know this shit's going to result in a ton of negative externalities and that the people who are going to have the most difficult time modulating are going to be young people whose brains are still kind of developing. And the question I would have is, do you want it to be like cigarettes, like opiates, like phones, and wait for 20 years of havoc and death and disease and disability and social unrest until we do something, or do we want to get ahead of the curve here? I've sort of come to expect to have my heart broken around this stuff, that every time we get close, nothing happens and that money wins here. So I'm somewhat cynical, but I want to still keep trying. And the EU here has actually kind of been a leader on these issues. But any closing thoughts on this one, Greg?
- [07:47]
B
I don't think it's 20 years, by the way. I'd like to talk a little bit more about what about energy costs? I think in some states we're going to be facing 20% higher energy costs kind of immediately now this year. Or how about job loss? I think there's a little bit too much focus on these longer term externalities that might take a while to really be seen. And how about what's happening right now? How did Elon Musk get to build a data center in Memphis with no sort of permit? So now it's off gassing noxious gas and sucking power and raising the cost of power in Memphis. We could have dealt with that. Now anyway, I would say one thing. Just vote with your wallet. Use anthropic. Their AI is called Claude and or do what I do, which is we will not pay for meta or Xai at our companies and we recommend clients not to pay for those AIs. They don't care about safety. And so at least we can vote a little bit with our wallets.
- [08:43]
A
Here I thought, what's his name? George Hinton, the father of AI. Is it George Hinton? Geoffrey Hinton, Yeah.
- [08:50]
B
Godfather. Yep.
- [08:51]
A
The godfather of AI. I thought he said a couple very interesting things that he's really freaked out. And his logic really struck me that there's never in the history of, of the world been a species that was controlled or not controlled by another species that was smarter. It's IQ that rules the world. Not if you have claws or how fast you are. It's whoever's smartest ultimately rules the world. And if you think at some point AI would be smarter than us, it's weird to think that for the first time in history that it would not control us. And the analogy I would use, that I've been using presentations because I think it's cool, is great white sharks are just much superior killing machines to orcas. They can breathe underwater, they have razor like teeth, their skin is thicker. But there's a couple of orca brothers in South Africa who coordinate and ram a dumber great white and then eat its liver. And basically every great white in Cape Town has either been killed or has just gotten the hell out of Dodge. And essentially these things, if they get smarter than us, despite all our weapons and all our sharp teeth and ability to breathe in different environments, it will figure out a way to dominate us. And his idea, Hinton's idea was to right from the get go try and build in a great deal of empathy into these models, empathies against humans. Sort of like that whole thing where robots are told the cyborg and aliens in the film Aliens, by the way, best sequel ever. Greg is told it can never. It's programmed such that it can never harm a human. It just is told that it can. You can never harm a humor. It's not possible for you to ever get to a decision tree where you harm a human. But this is going to be a huge question. I also wonder if. I think there's a non zero probability here that AI ends up being this really cool tool. But it's not. Doesn't have nearly the upside, nearly the economic benefit, and doesn't present nearly the kind of dystopian future we're all worried about that a lot of this is a little bit of like, I don't know, a little bit overdone and all these techno weirdos who like to think that they're more important than they are. Any thoughts?
- [11:00]
B
Well, I think the data is actually, I think indicating more of that than this than we're going to get. AGI. Right. If you look at. On the consumer side, the most popular use case for AI right now is conversations is either companionship or therapy or advice. Basically all a variation on I'm lonely, I need someone to talk to and get some advice. By far the number one use case on kind of personal AI. And people I don't think are really prepared to pay much for that. They want it. That's why GPT has almost a billion users, more than 10% of humanity. And on the business side, Scott, Enterprise AI is stalling out. We're now seeing data that shows that inside of companies, enterprise AI is getting to around 10%, 12% adoption. It's actually starting to flatline and decrease this past couple months. So I would say at this moment we're looking at data that would indicate this is an incredibly useful tool for consumers in particular, but they're not prepared to pay that much for it. And the jury's way out on whether enterprise can really harness these technologies and turn them into anything more than just a somewhat slight productivity boost. So you're right, this could end up being productivity software for companies and a kind of digital companion for consumers with not much of a business model. As I said, we'll know in a couple, three years and we'll see how this goes.
- [12:19]
A
So I too have heard that. I heard that the number one use of AI is actually therapy, which is essentially what you're saying. Look, you run an AI company that upskills professionals. For the enterprise. What are the primary uses in corporations for how they're trying to get their employees to leverage AI?
- [12:36]
B
Yeah, basically they are trying to automate away that bottom quarter of a team. The bottom quarter of a team are people that are basically cutting and pasting content and data and information. They're looking up information and passing it on. They're lubricants for data inside of companies, those roles. And that's what AI can do pretty well. And so I think this is the hope. I think that the challenge we have is that for the last three years, all employees have heard from the media is that they're coming for their jobs. AIs are coming for their jobs. And so this flatlining of adoption, I think has got a lot to do with, we don't know how to use these tools. And why would I? Because it's just going to replace me and take away my livelihood. So we've got to bust through this. This is a change management sort of behavioral change challenge, not a technology challenge right now.
- [13:35]
A
All right, on to question number two, which comes from Reddit. Interestingmilk37:77 who asks what kind of jobs will be most in demand over the next five to 10 years? And how long do you think it will take for businesses to switch most of their operations to AI? I get this question a lot, Greg. Any, any thoughts on where the ground zero is for employment destruction and which, which types of jobs are more immune or might accelerate in an AI world?
- [14:02]
B
So we know what ground zero is, and that's human translators, because those jobs have pretty much gone away overnight. And as we just talked about. Right. I think this has got to do with some intersection of how repetitive is your job and how much judgment do you use in your job. And if, and if you're high repetition and low judgment, low human judgment, then you're ripe for sort of AI disruption. The reality is they're going to be a bunch of new jobs created, but it's going to take a while. And there's only so many of those new jobs. Like, if you want to turn yourself into an AI person, I think you should go do that. You know, a prompt engineer, and now they call them context engineers. You know, you're going to find employment there. But for most of us, we're going to be in the job we're already in in a couple, three years. And we just need to be doing that job differently, using AI as much as we can. So I think about it as, first of all, just be in the top half of your team. Don't worry about go finding a new job. Worry about being at least in the top half of your team and the job you're currently doing. Yeah, sure. If you're a search engine optimizer as a marketer, you now need to become a generative engine optimizer as a marketer. And if you're so you're waiting for someone to ask you, train you to do that, you know, you shouldn't, you know, you need to figure that out yourself because your search engine optimization job is going to go away any, you know, soon and you need to know how to optimize for generative. But again, for most of us, be in the top half of your team, use AI every day and measure yourself on how many conversations you're having with AI. If you're having a couple a day, you're not AI enabled, you're not a super employee, you're a regular employee. If you're having 100 conversations a day with AI, you're probably a super employee. And those are the people we're going to keep on our teams.
- [15:46]
A
What I'm more comfortable projecting is that, okay, do you want to force your kids to take Mandarin in junior and senior year like they were doing of high school, like they were doing at tony prep schools in New York? No, that's just stupid. But I think the one place critical thinking on also just storytelling. Your ability to write well, your ability to craft a narrative, your ability to stand in front of people, your ability to create interesting content across the variety of channels that are out there. I do think that is the talent or the skill that endures trying to guess which industries or which jobs beyond the obvious ones get outsourced first. And the general, the general arc of all technology is that everyone catastrophizes about job destruction. There's some in the beginning, but usually over time the additional profits and margin created by that innovation creates new opportunities and new job growth. I'm still holding onto the notion that that's going to happen here. And I know, Greg, you think differently or you've said that may not be true. But I do think over time people are going to come up with so many different ways to use this that they'll start, there'll be new startups, new, you know, just a ton of new businesses leveraging AI to do interesting things for less money than they would have, such that you have different skill set and different people. I think there's just going to be all sorts of travel opportunities. I pay someone a lot of money to plan the travel for me and my family. That is going to have an agentic layer, an agent at some point so does she lose money? Yeah, but it will also create new jobs to develop those apps, maybe reduce the cost of my travel, meaning I'll be able to travel more, which will create new hotels, new travel offerings that will increase employment across other areas. I think that's the most optimistic I can be around this. So just some research here. According to the World Economic Forum, frontline jobs, including delivery drivers, farm workers and construction workers are expected to have the biggest growth. Nursing, caregiving and teaching roles are also expected to grow. I mean, I can't even imagine how many new electricians or construction workers or framers or people. All these data centers, I would think they're going to need hundreds of thousands of people just to build these data centers.
- [18:13]
B
Listen, I think, Scott, AI is truth serum at an individual level, at a team and organizational level. AI just reveals for knowledge work. AI reveals what's going on. Do you understand your inputs? What work do you do with those inputs? And then what are your outputs and how valuable are they? And I just say, I think this starts if you're a knowledge worker. This starts with sort of doing an honest assessment of your own job. How valuable are you really? What are your inputs? What work do you do to change those inputs into outputs? And could AI do it? Face this idea of AI as truth serum. Do it for your team. If you're a manager, how valuable is your team and how does it get work done? And will AI do it, improve it or replace it? Just kind of have that honest reckoning as fast as possible and then sort of make a plan. And if that plan is go become a plumber. Yeah, okay. That might be part of the plan. I don't think for most of us that's the answer. I think the answer is, you said it earlier, like, don't be mediocre.
- [19:11]
A
Yeah. AI is true serum. That's some real storytelling there. Is that how you kick off every meeting with a client? AI is truth serum.
- [19:17]
B
Do you like that? Yes, because I read that. Yeah. It's taken me 30 years, but I learned a couple of things from it.
- [19:22]
A
A few things. There we go. All right, we'll be right back after a quick break. Support for the show comes from Square. For most services these days, we're used to it being easy. Simply tap to pay. Think of your favorite food truck or a watch repair shop or a pop up art gallery. If you've ever had a swift, easy transaction with, with a business like that, there's a good chance they're using square. Square is a tool built for the way people actually run their businesses. It doesn't matter if you're at one location or multiple, or if you want to take your business on the road or online. With Square, you can take payments, manage inventory, run payroll, and send invoices. You can track it all from one place. It even includes the hardware you need to take payments in person and on the go. And it's designed to fit seamlessly into your space. Plus, the software is straightforward and intuitive, making it easy for anyone to run a smooth operation no matter your business background. Just set up Square and start taking payments quickly. With Square, you get all the tools to run your business with none of the contracts or complexity. And why wait? Right now you can get up to $200 off square hardware at square.com goprofg that's sq U-A-R-E.com geoprovg run your business smarter with Square. Get started today. Support for the show comes from Framer. On the one hand, you shouldn't judge a book by its cover. On the other hand, we all know what it's like to stumble on a janky website and assume that the operation behind it is janky as well. Bottom line, if you run a business you need a website. And if you need a website, you need Framer. Framer is the design first. No code website that lets anyone ship a product production ready site in minutes. That means that you have the freedom to build a website that is professional, polished and uniquely yours. No code and no compromises. Built in AI will handle the heavy lifting by generating starter layouts and the coding behind the scenes. You'll be able to a b test your design, set up funnels and see exactly where people click all from one place. And once you're ready to publish, Framer handles hosting blazing fast load times and SEO while you sit back, put your feet up and do literally anything else with your time. Ready to build a site that looks hand coded without hiring a developer? Launch your new site for free@framer.com and use code profg to get your first month of pro on the house. That's framer.com, promo code Prof. G framer.com, promo code profg Rules and restrictions apply. Support for the show comes from LinkedIn. We say this all the time on our show, but it bears repeating. Running a small business isn't just a full time job. It's about a dozen full time jobs that you can rarely if ever get to clock out of. At least until you get to the point where you can start Hiring the dream team. And if you've made it that far, you already know there's no time to mess around. That's where LinkedIn Jobs comes in. LinkedIn makes it easy to post your job for free, share with your network, and get qualified candidates that you can manage all in one place. And LinkedIn's new AI feature can even help you write job descriptions and then quickly get it in front of the right people with deep candidate insights. And if you decide you want to go the extra mile, find the perfect candidate. LinkedIn says that promoted jobs get three times the number of qualified applicants. It's all these little things that let you find help fast without compromising on quality, which add up to you finally having extra time in the day for, I don't know, relaxing or knowing my listeners, you'll probably use that extra time to expand your empire even further. Post your job for free@LinkedIn.com Prof. That's LinkedIn.com Prof. To post your job for free.
- [22:51]
B
Free.
- [22:51]
A
Terms and conditions apply. Welcome back on to our final question from rnrboy13 on Reddit. I work in product management strategy. I love using AI to speed up what I'm doing. We are encouraged to have AI write strategy documents in future descriptions. How does the company think about ownership and accountability when AI is making decisions? We're encouraged to have AI create the strategy for us. If that fails, is the AI to blame? I think that's an easy one. What are your thoughts, Greg?
- [23:25]
B
Well, we're not at AGI yet, so they're not smarter than us, especially as it relates to these kinds of decisions. I mean, for me, the fastest way to become irrelevant is to tell your boss that AI did it for you. Or this is the recommendation of AI. In fact, at Section, we're pretty clear about that. We have zero tolerance and people have said it in meetings, you know, well, this is what AI suggests. The job of AI is to get you ready to make the decision. And then you as a human has to make the decision. You have to own it and have conviction about whatever that recommendation or strategy decision is. So again, AI is your intern here. AI is not your management consultant. And if you offload AI to make these kinds of decisions, and I think frankly, you're not going to get a great decision anyway from AI. AI doesn't have the context, it doesn't have the real time knowledge, typically doesn't have all the kind of background information information that you have. So you're not going to get a great strategy recommendation from an AI in my opinion. And then second of all, we're not paying you to forward AI strategy recommendation. We're paying you to come up with one. So I think we're pretty clear on this one.
- [24:30]
A
Look, if you want to use AI, fine. To what extent you want to use AI is up to you. But at the end of the day, you're responsible for what you say. It's like saying, well, I know I was wrong and the ramifications were terrible, but. But Google said this and it really is amazing. I'll get responses back from queries on AI and I'll just immediately see, like, wait, this isn't right. Can you double check this? What's interesting is it goes, oh, thank you. You're right. And I want to say, well, why didn't you get it right the first time if you know, it's if a second question. That's what I don't get. As intelligent as it is, it does admit it's wrong. It's not like many of our politicians. It doesn't double down and go, no, no, no, the S and P is down this year. Trust me. It's like say, oh, yeah, you're right. Thanks for checking. And it gives you the right information. But yeah, look, you put your name on it, you present something, it's on you to double check every source and every technology that's helped you get to that point.
- [25:27]
B
Yeah. As you know, Scott, we're raising capital section our series B. And rather than ask AI to do the series B deck, which I could have done again, I would have got a very generic, not tuned fundraising deck. What I did is, you know, we did the deck. The humans did the deck. And then because you told me that the deck. The deck was too Canadian, that I wasn't being, you know, ambitious enough, I asked AI to take the Canadian ness out of our series B fundraising deck. And that's a great use case for AI and it did it well. So when you see the deck, you'll. You'll see that there's no Canadian ness in it, thanks to AI.
- [26:05]
A
Greg Shove is the CEO of Section, a company that helps deploy AI for enterprises and a good friend for 30 years. Greg, very much appreciate your time today.
- [26:18]
B
Thank you, Scott.
- [26:20]
A
This episode was produced by Jennifer Sanchez. Our assistant producer is Laura Gennar. Drew Burrows is our technical director. Thank you for listening to the ProPG pod from ProPG Media.