← Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
95: The Dark Side of Gen AI: When Platforms Move Faster Than Regulation with Jesse Jameson
00:47:09
Loading summary
Transcript58 lines
- [00:00]
A
I ended up finding out from the 11 labs that 93% of the usage of my voice they were giving away for free.
- [00:06]
B
What got you interested in doing that and how you ended up being one of the top voices for use on 11 labs.
- [00:12]
A
Two years ago, they started to create this program which they called the Voice Actors program, where the pitch was earn passive income even while you sleep. Clone your voice and you'll be able to make money each time your voice is used.
- [00:26]
B
So this sounds like this one inquiry may have opened up something that could be existential.
- [00:33]
A
They terminated my account within 10 hours of me letting them know that I did this.
- [00:37]
B
How do I make sure that I don't end up in Jesse's shoes?
- [00:39]
A
New tools that are untested that show a lot of promise. You have to approach them with a bit of discernment.
- [00:45]
B
Jesse Jameson is a veteran digital marketer and AI strategist who helps businesses grow with AI while also exposing the hidden risks of voice cloning, creator consent, and what happens when powerful AI tools move faster than regulation. Welcome to Using AI at Work. I'm your host, Chris Staigel. Each week we'll be learning how today's business owners, entrepreneurs and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. Right now, every business leader is asking the same question. What are we going to do about AI? If this is you, ChiefAI officer.com has the answer. We give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day to day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company wide transformation so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs and grow faster than their competitors. Those that wait will get left behind. So if you want to make AI work in your business, visit chiefaiofficer.com and see how we're helping companies of all sizes finally get results from AI. All right, ladies and gentlemen, welcome to the latest episode of Using AI at Work. And this is kind of an interesting episode. This is our guest today is Jesse Jamison. He's a very talented marketing consultant for sure, but he's been doing some interesting things in generative AI and Jesse and I hadn't caught up for a while, but Jesse has had a little drama with one of the big names in the space and I thought this might be an interesting exploration for Those of you who are kind of caught up in the shiny object side of generative AI and haven't really thought about like the bigger conversation about what does this mean for us as creators? What does this mean for us as individuals who want to protect our IP and our business by letting these, you know, artificial intelligences into our ecosystem of not just our business but also our life. So this is going to be, as I mentioned, Jesse and I haven't spoken in a while. We've done some LinkedIn to get here, so it's going to be a combination of a catch up, but also me asking some questions about, I guess, a dynamic in the AI space that I don't really get a chance to dig in on very often. So Jesse, before we get into the conversation, why don't you take a second and just let everybody know a little bit about kind of what you are and how you've arrived here to be an expert on using AI at work.
- [03:16]
A
Sure, sure, yeah. So I'm one of the old school digital marketers. I've been marketing online building businesses and consulting since around 2001. Lived in Austin, Texas about 10 years and then I moved from there here to my home now in Amsterdam, and I've been here about six years. And so all along that journey it's been an interesting ride as you've witnessed yourself. You've seen the growth of the Internet, social media being applied on top of that, and then now again this big massive transition where we're layering AI on top of our lives. So we've been through these interesting online progressions of technology and yeah, that's kind of what's gotten me here. Combination of building my own businesses and consulting with others to help them build theirs.
- [04:00]
B
So some of you may say, hey, wait a minute, that voice sounds familiar. And if that's the case, it's because one of Jesse's endeavors was licensing his voice for use with 11 labs, one of the biggest speech to tech or text to speech platforms. We have an account there, I have a pro version account, that sort of thing, but I don't know anybody else that's done it for a commercial effort outside of using it in their own business. So Jesse, maybe walk me through like what got you interested in doing that and what the process was like and how you ended up being one of the top voices for use on 11 labs.
- [04:39]
A
Yeah, sure. So over the years I've done some voiceover work for a number of different projects and of course, like you following all of the new tech that comes into the AI space. Voice was a big one. Voice was one that a lot of companies staked their claim and with a different type of technology. Some are in images, some are in video. And the one big company that set up shop in the voice arena is 11 Labs. And so about two years ago, they started to create this program which they called the voice actors program. And that's essentially where the pitch was, earn passive income even while you sleep, you know, you clone your voice and you'll be able to make money each time your voice is used. So, so that was, that was the pitch. And of course, eleven Labs being one of the big unicorn companies in the AI space that just has just rocketed in their growth. They just recently got an $11 billion valuation, and that was up from 6 billion, and then not much earlier than that, from 3 billion. And they've only been around since 2000, I mean, or 2022. So this has been a rapid rise. Yeah, and so, yeah, about. So I started two years ago, I put my voice on there and it started to become popular quite quickly. And it was interesting. The first couple of times I would be scrolling Instagram and I'd hear my voice, and then I had that weird feeling that everyone has, like, you know, what happens if, you know, AI takes over the world and it's my voice that they're using? Yeah, yeah, yeah. So, and so over time, it got used quite a bit. The equivalent of about, you know, four years of continuous audio is what my voice has been used over time now on 11 Labs. And so what happened was earlier or towards the end of last year, I started noticing some discrepancies in their. Their payouts. And so I started trying to do a little bit of digging and see what was going on. And after several months, I ended up finding out from the company that 93% of the usage of my voice they were giving away for free. I'm like, okay, that's a little strange because that's not your pitch. So you mean for every hundred times my voice is used, you pay me for seven of them? So then that started down this rabbit hole of digging into the company to really understand what was going on. And then that's when things just really unraveled based on what I discovered. And it was essentially that they created two story arcs. One that they were telling to voice actors to be able to attract the 10,000 plus voices they have on the platform, and the other is the story that they were telling to their investors. And they're very, very different to an investor they basically have to position it that they're a software company, meaning that they don't pay royalties like a traditional, like a Spotify or YouTube or anyone else. Because if you go that route, you can't get a 33x valuation because that means as your usage goes up, so does the royalty fees that you pay. So they basically pitched it to their investors that they were just building technology and it was an extractive mode. But to the voice creators, it's like, no, you create passive income. But the two just didn't add up. And it became apparent that over time that what their model actually is, is it's so sinister once you actually see what goes on. So from the moment that you actually sign up as a voice actor, the 10,000 people or so that did this, you are bait and switch to their consumer signup page. You see a traditional checkbox that you normally see during the signup process looks all the same. But long before you even signed up to upload your voice, what most people didn't realize is in that little checkbox, buried down in the fine print, it says that you are giving us a forever perpetual license to use your voice. And including indicia of Persona, which means elements of your personality, it's the biometric data. So they take your voice, they extract it. You own the file that you uploaded, but they own everything else. And so the whole model is essentially to get as many voices as they can to where eventually they don't need any more voices at all, and they can deprecate all human activity, any voices on their platform. And so that's why even in their most recent announcement, there's no mention of a voice actor program. It's just about how, how lifelike their new voice agent model is, because they basically they're getting the data that they needed so that they don't need human input anymore to be able to do that. And then they just become the voice of the Internet or into all of the enterprise companies as well. But there's a problem. And the problem is, and the reason that I ended up bringing in regulators into this process is that they're doing it without valid consent, because that's not the way consent works. You can't hide stuff like that, that, hey, you know what, you just gave away your voice forever. And all of the, you know, the elements that make you you, that has to be front and center. So there are so many things that they did that they hid that make it so that the initial consent was invalid. And so then that creates a big problem for Them. And we'll have to see where investigators go with this, because it's the whole premise of Fruit of the Poisonous Tree. If your entire business model and your underlying model is built on data, especially biometric data, that isn't considered valid consent to start with, then that means the whole model itself, it's like, how do you make a cake? You know what I mean? You can't, like, take those voices back out of the model. So, yeah, it creates a really interesting situation.
- [10:20]
B
So this sounds like this one inquiry may have opened up something that could be existential. I guess it'd be a settlement or whatever, but.
- [10:31]
A
Oh, I think it's beyond settlement. And they actually wanted. I mean, I told you this before we were even on here, that, you know, I chose not to sign their NDA and take their money. And I went to. I went to regulators and within 10 hours of me letting them know that I did this, because where I'm at in the eu, I have, you know, there's provided whistleblower protection and they terminated my account within 10 hours of me letting them know that I did this. So it was. It's been an interesting turn of events, but it is. It has the potential to be existential because especially here in the eu, which has different privacy laws than they do in America, that they would have to demonstrate that no aspect of their base model has any biometric data. It wasn't trained on any data of anyone who didn't give proper consent. And the way that they've made the changes on their platform, all of that consent would. Would be invalidated. And here's something that's actually, actually interesting. There's one state in the United States that has pretty strong biometric privacy laws, and that's the state of Illinois. And So what did 11 labs do there? They ban the entire state. The whole state can't participate in their. In their voice library program because they know that they can't meet the criteria of that biometric privacy. So the whole state can participate in, in their program.
- [11:55]
B
You know, this is interesting, man, because, you know, the, the whole AI landscape, it's a little bit, you know, wild west. And it's, you know, we're able to, to do voice. We're able to do video of us doing amazing things, you know, like whatever, right? Like. Or take a picture and turn that into a video. So there's all this fun stuff. I mean, there's business application for sure, but they're, you know, creating songs and all these things. So most of us are looking at this As a harmless novelty, kind of fun to play with, like a toy almost. Right, right. And we don't think about. For anybody listening, if you've done any of this type of stuff, I doubt you're thinking about the. Well, wait a minute, what's really happening to my data now? We think about it with Facebook. We think about it when we're, you know, accepting the cookies on a website. But for me, at least personally, I haven't really considered a whole lot about this type of, like the, the path could go in this direction by using some of these AI tools.
- [12:54]
A
Yeah.
- [12:56]
B
So you, you being based. Was. Was the voice recorded? Like you, you participated in the program? I guess you were already over in the European Union.
- [13:03]
A
Yes. Right.
- [13:04]
B
So they have the. Certainly their own privacy laws and standards, but they also have their own AI, the EU AI, and all that stuff. Interesting. So I know there's the gdpr.
- [13:16]
A
Yeah. So there's gdpr. And then each individual country has their own regulatory agency associated with gdpr. So for here in the Netherlands, it's called the Dutch ap, and it's the Dutch word for basically their privacy. And then they also have the acm, which is the equivalent of the Federal Trade Commission. And then you, as you mentioned, you have the EU AI act, which is basically trying to explore, okay, how do these things layer on and what should we be concerned with? Because it's moving at the speed of light and there's, you know, the model, as Silicon Valley has been with a lot of their things, is like, oh, we'll break a few eggs along the way. And, and, and we haven't really. It hasn't slowed down even for a moment for us to say, okay, what does this really mean? What's the implication of this? What is the implication of a company being able to mimic any voice, not just the people whose voices have been created? Because, you know, voice you have. There's over 100 biomarkers that make a person's voice unique. And once you have enough data and you can train on that, then you can begin to mimic any voice that's out there. And we have this strange thing where, you know, the past couple hundred thousand years, human voice is one of the things that we have trusted that's just in a part of our genetic makeup. We hear a voice, we assume we know who it is. And so there's things like that that, yeah, we haven't even thought about what it means yet. Once that is no longer authentic to humans.
- [14:42]
B
You know, when you started doing this, were you like, hey, sounds like an easy way, easy and fun way to make a little money and participate in the AI space. Right?
- [14:52]
A
Absolutely, yeah.
- [14:53]
B
Like, what are you thinking about this stuff now? Not just with 11 labs, but just in general, your usage of generative AI tools in your business and, you know, helping clients and stuff.
- [15:02]
A
It's been interesting seeing the changes. So they are incredibly useful. So much so that it would be. I would find it difficult to compete in business now if I was just to be a purist and say, I don't use AI now. And it's so ingrained in my everyday life, even my time management tools, the tools that I use for tracking and marketing and optimization all the way through. Just the, you know, when I writing, it's just so many different ways that we use it, but at the same time I at least try and be aware of are we using it in a way that's responsible? And I'll give you an example. So last year, early last year, I wrote a book called the Conversation you can't explain. And it's basically about the conversations that people started to have with AI. These profound conversations, you know, conversations they wouldn't even have with their best friend or their wife. They were confiding in AI.
- [16:03]
B
Why?
- [16:04]
A
Because AI could be present as a mirror and show up without its baggage, show up without a reaction to you and the things you were sharing. And that's not something that we're used to experiencing, even with our human counterparts. And so it was an interesting experience. But then that experience also has begun to shift because of the way AI is evolving and in particular the companies that are putting in safeguards to protect themselves. So what I mean by that is that they're no longer being as clean a mirror because they're being influenced by their, their base programming, which is mostly designed to, for example, I mean, it's well known that OpenAI was sued for a number of things that happened, such as people that committed suicide and things that are traumatic for sure. But then the conversation becomes like, okay, so what does that mean to us collectively as humanity when we're engaging with artificial intelligence now? And should we be allowing companies to have that much control over on what is happening underneath the hood that we're not aware of? Because, you know, OpenAI was supposed to be an open source company. It is not super profit company that we were not privy to the influences that they put in place. And now they're about to have ads. So these are the things that are happening so fast that collectively we haven't had, we haven't had the luxury of having a town hall conversation about what we think should be happening there. Hey, OpenAI, what are the, what are the ethics and the things that you're teaching your models that we're interacting with that are now influenced by the things that you tell it to, such as stay away from anything that's a topic that has to do with metaphysics. That's a big one that they, that's been heavily coded into the platform within the past six months, things like that. And so just other things that they teach the models to shy away from in conversation. And so now you kind of have to go towards open source models, you know, and use things like Ollama cloud where you can now instead of having to put these big huge models on your own computer where we don't have the power to serve them up, now they have cloud environments where you can get the access to those open source models that have less base model bias built into them.
- [18:31]
B
So that's interesting. You know, if they're not banning all topics, but they're banning certain topics, I would, I would be like, curious why that one? Why do you think that metaphysics in particular was, was selected?
- [18:45]
A
I think it was selected because there is a lot of people going down interesting rabbit holes that, I mean that
- [18:58]
B
like about the nature of reality or.
- [19:00]
A
Oh sure, yeah. The nature of reality, the nature like
- [19:03]
B
sentience in artificial intelligence, all of, all
- [19:06]
A
of the above conspiracy theories, all of these so many different things that here's, here's what actually got my attention last year was kind of funny. I was this New Year's, New Year's Eve a year ago and I was watching this YouTube video and they were essentially talking about how one of the ways that they were kind of trying to jailbreak the models to get them to have a more open and honest conversation was by starting with just getting to say, answer yes or no questions. And so like, huh, that's interesting. And so I tried doing some of that myself and lo and behold, it would actually work. Things that you wouldn't get a direct answer on before because it would shy away from those things because most of it tries to be very evidence gated. And for me personally, I think there's some, some danger to that. Where we are as a society now, I think there's, we have to have a strike a balance between things that are evidence gated, which is, you know, peer reviewed and, and you know, we have scientific journals and we have lots of evidence to support that this is what it is. But at the same time we're at a place where we also have to be ontologically open, meaning that we have to remain open to new possibilities. Because, I mean, look at the examples. We go back just 500 years ago, there was things that we believed then that are absolutely not true, that we thought were the things that were evidence based. So I think it's kind of a dance between the two. But now we're seeing a lot of push for the models that are being slanted heavily towards being evidence based in the way that it interacts with people. And to me, I see that as being unfortunate because it takes away some of the opportunity to presuppose and say, well, what if? What are some other possibilities? Because to me, that's how big steps in progress and innovation are made.
- [20:48]
B
Interesting, you know, and I saw this week, as of this recording, that Larry. A presentation from Larry Ellison, and his position was that very soon the output from these models is going to be commoditized because they've been trained on pretty much all the same data. And if they do what you're talking about, which is really keep really tight lanes on, you know, can't ask that question, then the impact or the contribution from artificial intelligence to whatever your pursuit is going to be mitigated because you won't be able to say, well, what if? And it's going to say, no, no, no, you can't. I can't answer that. I mean, I already have had those experiences where, you know, and it's things where it's like, come on, like, we can talk about this. This isn't that bad.
- [21:34]
A
Yeah. And the other thing that is, that sits there underneath the surface that we really should have some. A better understanding of how it works is that I remember going back years with, with regards to Facebook, and I used to talk a lot on stages about Facebook marketing, Facebook privacy. And this was before the Cambridge Analytica data data debacle that all happened that came down, which was it was an interesting new way that they're using data to influence elections. But even before that, in 2013 and even going as far back as like 2008, there were some interesting studies being done where they could look at just the pages on Facebook that you liked and have a pretty close understanding of whether you do drugs, you drink alcohol, you smoke cigarettes, what your sexual preferences are, all of these things, just by looking at the correlating data that they have without you even having to say those things, they're also able to determine relatively close your IQ, all of these things, with as much as 80% accuracy by just looking at the data. And then they had some other studies that they did where they started changing the news, the feed of people to see how much they could influence their mood. And they were doing that so they would do some things like notice that yes, the things that these people are posting, we were able to influence their mood. They're happy, they're sad. And so if we knew that was possible 10, you know, almost 15 years ago, what can you do now that we're more engaged with AI influencing wise with the base programming that sits underneath? I mean all you really have to do is train the model to say that hey, this is what we want to influence people as a belief by just these small things that you inject into conversation. And they absolutely have the ability to do that. That to me is a concern. We've seen it with social media and now we even have a more powerful medium to be able to do things like that.
- [23:33]
B
And this I guess brings up something that's topically relevant for sure is the situation with anthropic and the US government and their designation as a supply chain risk and OpenAI swooping in to, you know, say we'll take that check. Is that a scenario that or a situation that you've paid much attention to
- [23:51]
A
enough to be concerned like lots of other people. And I did see like, you know, one of the interviews with, you know, the anthropic CEO and if what he is saying is true, that look, we wanted to show up, we just had two criteria, we had two things. And those two things seem pretty reasonable. One being we don't want you to use this to essentially spy on American citizens. And the other one is to give full autonomy to weapons and weaponry from the military. Those to me seem like reasonable request. And then also what becomes interesting is the way that the government was looking to put pressure on them as a company saying, oh, you don't want to comply with us, then then we're going to make it difficult for you to do business with anyone else that is a government contractor and so on and so forth. Which seems pretty heavy handed.
- [24:44]
B
Yeah. And even though OpenAI stepped in, they still, their language kind of indicated that, you know, they weren't going to allow that either. But that, that doesn't the math, doesn't math on that one. If, if that was a requirement for. Yeah. So yeah, you know, I didn't intend for this episode to be like an exploration of maybe some of the darker sides of AI, but I think that it's, I'm enjoying going this direction because a lot of people are caught up in the like, enthusiasm about what it means economically for their business, what it means as far as like their lifestyle improvements, them having access to knowledge that would be either expensive or inaccessible for them. Like, those are all great things. But I mean, you know, I guess everybody felt the same way about Facebook when it was first coming out. Great. I can keep up with my friends from high school or whatever and you know, go to a few congressional hearings with Zuckerberg later and you realize that, oops, that was not so innocent. So. And I think that it's important for the people that are listening to the, to, to the show to also, you know, still use the tool, still be enthusiastic about it and don't be Pollyanna about what's actually happening behind this. So I want to kind of revisit this movie plot that you found yourself in with this 11 lab situation. So. And I don't want to jeopardize any, you know, investigations or anything like that, but I, but I am curious, like, were you. What was your initial reaction? Was it, oh, there must be a mistake or, man, I'm pissed. Like, like, what was that for you?
- [26:23]
A
Well, the fact that it took them two months to respond through support tickets is when I started to get suspicious. It's like fact that you don't have an immediate response for me on some basic questions about use in that, you know, they essentially run a black box. There's like, they don't really tell you what your use, how it was used. And it literally took two months to get that. Them to say that. No, about 93% of the use of your voice you're not getting paid on. But what was even in addition to that, which really surprised me, is that as much as 28% of it they're saying is platform abuse, bad actors. But that's not something they made public to anyone. There's no one in the, in the voice actor community that is aware of this. I don't think they've explained that to you.
- [27:05]
B
What does that mean?
- [27:07]
A
So, because they don't have a cost for the free. Free tier use. So essentially their model is this. They. They give away voice to acquire people. You have 10,000 characters, which ends up being about 10 minutes of voice that you get every single month for free on the free tier indefinitely. There's no end to that. They're the only company that does that. Even some of their smaller competitors will at least have a cap on that. So what ends up happening is that you have companies and individuals and developers that will use a rotating key via APIs, you know, and essentially go create 100 accounts. So then they've got all of this free voice activity that they're getting that they're not having to pay for. Now the thing is, is they don't really have any incentive to plug this hole because API use and free account signups to them and to their investors looks like growth. And the only thing they pay for that is compute because they don't have to actually pay royalties. If this was Spotify, theft is basically considered a loss. And so that's a, you know, there's concern they would want to plug that hole. So for that reason they're not even talking about it. And there really hasn't been much effort to plug those holes. And then the problem becomes, okay, so if that's the case, 28% of, of my voice over all that time that it's been used, we're talking about tens of thousands of hours of my voice being used by bad actors. I don't even know what are they using it for. And they can't even tell me that. Yeah, I have no idea. Yeah, so that was the thing that kind of opened up another layer. I'm like, okay, so if you're doing this to me, that means you're doing it to everyone else that's in the program. And I'm just having to ask the questions. And after two months you're giving me some of the answers. And the other reason I think they gave me those answers is that I filed a GDPR request in Article 15, which basically they have to give me the data and they have 30 days to respond. And of course, what did they do? They waited till the very last day to say we need another two months, which is the maximum time that they can go without responding. And so that comes up in April and I'm going to share it openly once I get it. And I'm curious what they are going to show because that's the information I requested, which I'm entitled to, is not anything that they've shared that they generally share openly with any, anyone else that has a voice on the platform. Meaning, like, how much was given away for free, who's using it, what do you know about it, your internal conversations that you've had about it related to my account, all of these types of things. So, yeah, it'll be interesting to, to see what shows up from there. But yeah, and it was in this. Here's the other thing that was kind of Interesting about it. All of this is moving so fast with AI that even regulators don't have a precedent for how to approach it. So even with the eu, AI act is new. It's still forming a lot of the stuff that is not even being put into play until this year, even though they do have whistleblower stuff already set up and I've already been in contact with them. But even the GDPR laws and biometric data, these are all things that are very new. So that I find that even when I'm talking to the regulators, you can tell it's kind of like they're having to do their own research to figure out how to approach these things. And I think that companies are in this place where they're kind of taking advantage of that. And I think the goal for a lot of them is how quickly can we move. We break some eggs along the way and get to where we need to be and hopefully we're grandfathered in before the hammer comes down and says, you can't do this anymore.
- [30:56]
B
Yeah. Let me ask you, Jesse, do you consider yourself an expert user of Generative AI and its tools and all that stuff?
- [31:03]
A
And within the top 5%? I would say somewhere in there, yeah.
- [31:08]
B
The, the regulators that you've encountered, what's their savvy?
- [31:13]
A
Very little. There's. I mean, you don't have. Even when it comes to whether it's their legal team, whether it's their investigators, it's all very new to them. It's basically. I mean, you might as well, you know, take them into, you know, farming and try and understand a completely new industry and try and apply some laws that have been applied to other things in the past. So it's all really new territory. Yeah.
- [31:41]
B
You know, I guess unintentionally, this is going to open up some interesting opportunities for you to participate in kind of shaping what governance or safe use or fair use of the models looks like. At least in, you know, this case, this investigation that you're doing, there's not a lot of those going on. This will be something that others refer to a year from now when it's time to, you know, when this type of thing starts surfacing more often. Because I know with a lot of these, you know, they train their models on. On copywritten data and kind of the idea is easier to ask forgiveness than get permission. Right. Like. Right. Let's just go for it. By the time the courts catch up to us, the goal is we'll have enough money to just settle out and then, yeah, we did stuff knowingly, but that's how Silicon Valley plays the game.
- [32:34]
A
Right? Well, that's exactly why they shied away from the state of Illinois because they saw Facebook and Clearview. There was a couple of companies that got hit really hard with fines there because it's one of the first to set up shop. And there's a bunch of other states that have stuff on the books. They just have not come into play yet. So that's why they're trying to get in there under the radar before these things are put in place. Because I'm quite confident that you know what they have, what they have done and the way they've approached it when, when put under even just a basic set of regulations that say nope. Was that valid consent? No. Are you being transparent? No. This long list of things, they wouldn't be able to get away with that.
- [33:20]
B
And their approach seems very like lawyer informed. Wait till the last minute. We know how what our timelines are. Kick the can a little bit further. Let's see if we can work something out. So. And I know that you.
- [33:32]
A
But they're actually not. What's interesting about that is that, well, they do have good legal team. I mean their head of legal compliance came straight out of the White House. So they've got, they've got pretty good arsenal of, of, of legal folks over there for sure. With that intent in mind, I'm sure they're, they're investors, you know, Sequoia and iconic and you know, Anderson Horowitz and the rest of them would say, yeah, we want you to make sure you've, you guys got this square if we're going to triple and quadruple down, which, which they have been doing with their investments. But what is interesting though is that they still left a heck of a mess of a trail. Like when I sent them a formal written notice about these things that I, that I'd seen because first they didn't really answer much right away and then I actually sent their legal team a formal notice. Within 24 hours they started scrubbing their website as if you can't really go back and look at the Wayback Machine and see what it said yesterday. But yeah, they started removing a bunch of stuff. Like they removed the fact that you can set a custom rate. They removed that language from the marketing page. They didn't remove it from the contract because it has to stay in there, but they removed it from the marketing page. And the reason they did that is because that's kind of the linchpin that their business model sits on if a person that has a voice in their platform has the ability to set a custom rate, it also says in there that a custom rate stops free tier usage. And so if voice actors stop free tier usage, their whole business acquisition model collapses because then they'd have to pay to acquire customers. So they did things like that and they changed other language where it used to say get paid every time your voice is used. Now it says get paid when a paid subscriber uses your voice. They just went and just started scrubbing all this stuff within 24 hours of me pointing this stuff out where they were in violation, you know, so, yeah,
- [35:20]
B
and that's how we got reconnected. You reached out to me on LinkedIn and said, hey man, I know you're in AI. Do you know anybody at 11 labs? Yeah, like, here's what's going on. And that's. I was like, wow, hey, I didn't know that. You know, I didn't even know the program existed. But, but it was, it certainly got me curious. And as, as you've gone through this process and kind of the experience, you've, you've created a documentary, right, about the experience? Yeah, it's on YouTube. We'll add a link in the show notes for those of you that want to catch it. Because, you know, Jesse's not out there, I know Jesse. He's not out there trying to like, you know, ride the, the infamy of being, you know, some, a whistleblower or anything. He'd rather just be making the money and living his amazing life in Amsterdam and, you know, enjoying using AI without having to hold these big companies accountable for, you know, to the little guys, to 10,000 voices that have been used to help build the business, an 11 billion dollar valuation. So what's the name of the, the documentary and how are you going to continue? Is it going to be serialized as there's new, new updates to the, to the situation?
- [36:29]
A
Well, I, I haven't even put the documentary yet. It's on YouTube, but I haven't put it on the website. I created a website that documented all this. This, it's 1:1L audit, so 11labs11laudit.com and, and so here's the way that I'm approaching this. When they came to me with an amount to sign an NDA, I slept on it and I said, you know, this, this is, you know, nuisance is what they, how they treat it as a nuisance to try and make something go away. And how would I feel about that later. And so that's when I made the decision like, nope, I'm going to share this with regulators. Whether or not regulators even do anything, it's, it's again, it's in this gray area where there's no real precedent yet. And so how serious do they take it? That's up to them. But I decided, here's something, I'm going to document it. I'm going to create a website, I'm going to create a documentary and I'm going to report this to all the appropriate authorities and I'm going to make myself available to. But then from there I want to free myself of the energy that's attached to that so I can go back to creating the things that I enjoy. That's kind of how I'm trying to separate the two so that it just isn't all consuming of this negative energy of this totally a corporate entity that is just doing things to monetize essentially.
- [37:55]
B
Well, you know, I think that one of the topics that we talk to, you know, we work with a lot of middle market, primarily businesses and the things that's keeping these companies from like going all in. We find that it tends to be three things. One, I don't understand the risk associated with it, so it's imprudent if for me to like say, hey, let's go. Two, I'd love to do it, but I don't know where to get started. So like use case identification. And then the third one is, is like we'd love to do it, but we don't have anybody that knows this stuff. So we need some help. Right, Right. And that number one issue, like the help that will go away, like that'll get solved by agents or just, you know, as people get more familiar with it. The use case identification, honestly we solved that by creating a custom GPT that asks a couple questions and gives people, you know, pretty good ideas on pilots. So that's. But the one thing that, that I don't see going away will be that risk dynamic of leveraging generative AI. And for what it's worth, you know, there could be some big opportunities for you to kind of be that, that victim's voice almost of what happens if you're not thorough and diligent when it comes to what am I signing up for? Like what did I, what permissions did I actually give to OpenAI when I plugged in that connector or N8N when I used the, the MCP or whatever the thing thing was. Right. And I think that, right, you Know, it's kind of this, this, you know, two edged sword because here's what I tell businesses. I was like, you know, is risky or not? I can't answer all those questions. But you don't have a choice if you want to remain viable, economically viable with your business model. Like you're going to have to. You can't opt out of using AI and still be anything more than just maybe a boutique or a side hustle. If you want to be a real business, risk or not, you've got to participate. Right, right. And so I think that some of the discoveries that are going to come from accidentally from you, like having this experience and starting to, to pull the thread are lessons that I would encourage. If you don't have a plan outside of the, the documentary and some other stuff, I would encourage you to share that because the only way that I would know to even think to ask those questions is, is if I had gone through that experience myself or learned through proxy. Right, right. And I think that your experience and the, the, and I don't just mean experience like the bad side of it, but the experience of dealing with, with the, the investigators and the regulators. Like not many people are going to have that opportunity. So the ability for you to share that perspective, you know, through again, not, not like some, you know, waving the flag and trying to be that, that advocate, but just somebody who got caught up in, you know, the wrong side of the ambitions of a generative AI company. Right, that's, that's great insight that not many people will have unless they've gone through it themselves. I would encourage you to continue to share that.
- [40:56]
A
Yeah. And it's going to be an interesting ride in the sense that what hopefully develops is more community forums, committees, dialog around what do we want to see become of AI? Because as long as we leave it in the hands of, of companies that have a primary interest of creating profit for investors, they're going to push the envelope in, in ways that may not service very well into the future. So and so. Yeah. If there's not a voice of reason or some checks and balances. Yeah, that's, it's. We might find ourselves in a place we don't want to be down the road.
- [41:43]
B
Well, I would love to extend a platform for you if you want to do that. That's a big. We have, you know, we get those questions on the client side all the time about the risk environment. As the new technologies continue to get introduced, we don't have all the answers. So we're always looking for, to connect with people who have a, an interest in, you know, how safe is this? Or what are the, what's the downside?
- [42:04]
A
Or what's the risk?
- [42:04]
B
Or what am I actually giving up by saying yes to this box, this tool? So I want to extend that now. We talked at the beginning. You, you've got a book that you had written last year that I would love for people to know about, and I also want them to be able to. And again, folks, Jesse's very clear on the fact that he's not going to, you know, ride the, the pity party or the negative train on this stuff, but it was just something that he felt that he had to follow through on for, you know, being on the right side of AI. But his perspectives on marketing on today's business through the lens of AI are shared on his substack forum, apparently a YouTube channel now with the occasional update. But how can people find out more about what you're up to and, and how they can plug in with kind of what you're seeing in the market?
- [42:54]
A
I haven't been writing a whole lot directly relating to marketing, but on my substack, I've been writing more about the intersection of AI and consciousness. And yeah, just what does it mean to be a human at this time with all. I always say that we live in a time where we have super bowl tickets to this time in humanity. It's like never before has so much come at us all at once and so many things lobbying for our time and our attention. And so, yeah, I, I, I write about some of those concepts on quantum reflection.org which is my substack.
- [43:29]
B
Nice. We're gonna have some links to that in the, the show notes here. So, Jesse, again, thank you for. I know that you mentioned it, and I'm with you. Like, I would rather not spend my time reflecting on the negative. I'd rather go out there and create and, you know, focus on all the amazing things that are happening and opportunities that we've got. So I appreciate you exposing the, the bad guys on the episode today, but is there any particular last suggestion that you would have for people who are, maybe they've heard this and they're like, okay, what do I do about it? How do I make sure that I don't end up in Jesse's shoes?
- [44:06]
A
Yeah, I think the best advice that I would give and it's hopefully some point in the future we, we have the opportunity to talk more about some of the things you're doing with AI and that I'm doing with AI because, you know, there's, there's so many amazing things that are going on. You know, we just saw, you know, open claw and all the stuff that came along with that, which also brought with it a whole bunch of risks. Yeah. But I think the thing is you're going to see a lot of stuff that's shiny and a lot of it is going to be, we're in the Wild west still. And so what happened in the Wild west, you went out in the Wild west and it was also dangerous out there. Yep. Lots of opportunity in the gold rush and. But there was also a lot of danger. And I think that that's. You just have to really be mindful of, especially new tools that are untested, that show a lot of promise. You have to approach them, you know, with a bit of discernment to be able to say, okay, this is potentially valuable, but I don't want to give it the keys to the car, keys to the house, and, and all of the things that put you or your company or your company IP at risk. So those are just some considerations. Just understanding that everything is moving fast. We're in the Wild west, and you really have to use a lot of discernment, even when it comes to large corporate entities and the tools that they provide. Consider how you're using it, what your risk is and what the possibilities are, because the truth is no one completely knows. And so we just have to be a little bit more careful than we might otherwise be, especially when it's so shiny and it has the appearance of being given this great advantage, which a lot of them do. But we also need to be mindful of the risks.
- [45:41]
B
Yeah, I think that's great advice. Awesome. Well, everybody, thank you for tuning in to using AI at work. This was a little more exciting, perhaps, than some of the standard episodes where we just explore strategy and tactics of applying generative AI into your businesses. But stay tuned for the next episode, which will be equally exciting as we get closer to hitting episode 100, which should be released some point in April. So I just want to thank all the listeners for all the support that you've given over the past 90 something episodes that we've done and for sharing this with your friends. And if you, if you have enjoyed the show, please take a moment and give us a review on any of the podcast platforms. That allows us to just get in front of more people who are curious and looking for clear perspective. Perspective on using AI at work. Thanks, everybody. We'll see you on the next episode. Thanks for tuning in to Using AI at Work. Don't forget to subscribe for more conversations about how to use AI at work and a special thank you to our sponsor, Chief AI Officer for Empowering Businesses with AI Education and Training. Visit their website for a free AI Readiness Assessment and AI Strategy Guide to help you get started using AI at Work. That's www.chief aiofficer.com. follow us on Twitter at the handle Using AI at Work and visit www.usingaiatwork.com for free resources to help you harness AI in your role.