We Fixed It, You're Welcome - AI Privacy with Amazon Alexa: Control or Convenience?

Episode Date: April 8, 2025

In this episode of "We Fixed It. You're Welcome," the hosts tackle Amazon's AI assistant, Alexa, and its recent privacy policy changes. Joined by guest Christian Johnson of Metis Analytics, they explore the implications of Amazon's data collection practices and the future of AI in our homes. The discussion covers the balance between technological advancement and personal privacy, with perspectives ranging from embracing AI's benefits to concerns about data ownership and identity protection. Christian offers insights into edge computing and local AI models as potential solutions, while the hosts debate the ethical considerations of AI development and data usage. The episode concludes with recommendations for Amazon to prioritize transparency and user control in their AI offerings. https://www.metisos.co/Evolution of Amazon's AI Assistant Discussion of the AI assistant's development since 2014. Mention of over 600 million devices sold and initial financial losses Privacy Concerns and Data Collection Recent rollback of privacy protections by Amazon. Introduction of OWL Plus and its implications for data collection. Transparency and User Control Importance of clear opt-out choices and positioning privacy as a feature. Comparison with other tech companies' approaches to data privacy. AI Development and Data Usage The need for data to train AI models and improve services. Balancing technological advancement with user privacy. Generational Perspectives on AI and Privacy Differing views on AI adoption across age groups. Discussion of the trade-offs between convenience and data protection. Edge Computing and Local AI Processing Potential for AI processing on local devices to enhance privacy. Christian Johnson's insights on edge servers and personal data control. Ethical Considerations in AI Development Debate on the use of personal data for AI training. Concerns about identity protection and data ownership. Future of AI and Personal Data Management Exploration of individual language learning models (LLMs). Potential for users to have greater control over their data. Amazon's Opportunity for Privacy Leadership Suggestions for Amazon to champion privacy and build trust. Importance of clear communication and user-controlled privacy settings. Balancing AI Advancement with Privacy Protection Discussion on the necessity of continuous data collection for AI improvement. Exploring alternatives to centralized data storage and processing. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info. Learn more about your ad choices. Visit megaphone.fm/adchoices

All right, here's how this works. In each episode, we pick a company we all know that has something going on right now. Then we put ourselves in charge and see if we can fix it. You'll be hearing from Melissa and Operations, Chino on people in culture, and me on marketing. My name's Aaron. As always, a quick disclaimer, we are going into this somewhat cold and nothing we say should be construed as legal advice, financial advice, or anything that would get us in trouble. These are our views and opinions. We're here to ask the kinds of questions everyone's thinking, have an engaging conversation, and maybe come to some conclusions that we feel are worth exploring. By the end, if we fixed it, you're welcome. All trademarks, IP, and brand elements discussed are property of their respective owners.

Welcome back to We Fix It, you're welcome. It's great to be here, great to have you here. Today we're going to talk about a certain innovation from Amazon, how it got from where it started to today, what the future holds for all of us. Spoiler alert, it impacts our data privacy. As I've a good thing? Should we put a stop to it? Can we fix this situation? We will find out. But not before we introduce today's esteemed guest, Christian Johnson of Medis Analytics. Christian, tell our listeners a little bit about yourself. First of all, thank you for having me. My name is Christian Johnson, founder of Medicine Analytics. We build agintic systems for small and medium businesses. Right now, people are just really learning what agentic systems really are, but just really think

of them as learning systems, systems that learn and adapt that can use real-time data like APIs and sensor data to be able to really interact with the world and really perform actions on the behalf of humans. So we build them. We're a full stack development team. So we have right now, I have a team of 10 people working full-time on legitic systems. And we love to build. So thank you. Thank you, Christian. All very pertinent to our conversation today. Thank you for joining us. So let's do it. Let's get into what we're here to talk about. This is the first episode where we can't actually say the name of what we're here to discuss. It's not out of censorship, but out of consideration to our listeners. We'll try to avoid the wake word of Amazon's smart assistant. Otherwise, we might wake her up and accidentally cancel your shopping lists or set off your alarms or any number of bad things. So instead of saying, A-L-E-X-A, let's all try to. follow Paul Simon's advice and call her Al. Does that work for all of you? Can we do that?

If we trip up, I'm sorry. So yes, we were talking about the AI assistant that Amazon launched over a decade ago in 2014 in the form of an echo speaker. The first echo dot smart speaker looked like a hockey puck, and it was impressive at the time, but pretty basic by today's standards, already a decade plus later. It could play music, set timers, tell you the weather, maybe answer a trivia question or two. And that was about the extent of it. And over time, the smart assistant in question evolved into something much more powerful. Today's devices control your entire smart home ecosystem that come natively installed with your TV and even integrate with generative AI, which I'm sure we'll talk more about, to hold more natural conversations.

We know that Amazon has said there's been over 600 million AL devices sold. We also know, according to a Wall Street Journal, that the first four years of rollout lost the company $25 billion. So it seems they've been playing the long game, which may be influencing the decision we're here to talk about, because it sounds like it's about time to monetize, which they've been doing, but maybe escalate because Amazon's a data-driven company. Over the past year, they've been quietly rolling back some of the privacy protections that have been in place since the beginning, now making it harder to delete data and defaulting more settings to allow data sharing. And we've got Al Plus rolling out, which was announced in February as a paid subscription that's free to prime members for now, which makes more use of AI, which we'll also talk about. But also, as of March 28th, if you use an echo smart speaker or an echo show, the one that has a screen, everything you say when you wake it up will be sent directly to Amazon for analysis and training. And there's no opt out. So what I want to know is, should we let not just a voice assistant, what we say go, no, go.

not just to the voice assistant, but to the company Amazon itself, should we let them listen to what we have to say unfiltered for the sake of better technology? Is this a good trade-off? Let's hear it. Well, I would have to say that I think really people lose trust in systems that aren't protecting your data and doing a really good job of that. One thing I've seen, even as I'm also an iOS developer, and one thing I really appreciate it about Apple is that they do a really good job. good job of that and making sure people, they really respect customers' privacy. So I really think

that, you know, large, big tech really has to be staying focused on privacy. I think, Christian, you bring up a really good point. So one of the things that I think is very key here, Aaron, is that we are talking about transparency and understanding the policy, right? So understanding what the data is being used for when you're being listened to by multiple devices. We all know that that's happening around the world. It's interesting because, Christian, you bring up Apple, but they came under fire at the beginning, right, when it was revealed that they had contractors that were reviewing Siri recordings without our users' explicit concerns.

And so that backlash made them really stop and pause the program and rethink how they could utilize those recordings because most people might not really care, but give it in a more opt-in scenario, allowing the customer to feel like they are in control of their own data. It also allows users to delete past interactions. And so I think Google's doing the same thing where you can disable or clear your history. And so really kind of a lesson for Amazon around of this would be around prioritizing the privacy aspect as a competitive advantage, learning from Apple, Google, meta, all the folks that have kind of stumbled through this, and using more clear opt-out choices and position privacy as a feature, not as something that they're using and that

they don't really care about what it means to you personally. So that's kind of my thoughts on that as well. So just building upon what you have to say. Thank you. Amazon does mention privacy on their release announcement for Plus, but it's about the seventh or sixth or seventh or eighth point made. It's pretty far down the page saying, of course we care about your privacy,

but they're not leading with it. And, you know, is that important? Are people paying attention to the privacy changes and what's happening right now is, you know, is what is the reason for deemphasizing it compared to all the other, you know, look what you get. But here's a little bit about your privacy. I think it's quite interesting from kind of a user perspective, right? Obviously, it's been just over a decade of these type of devices coming into our homes. And I remember at the time, it was exciting. And you

could talk to Google or Amazon or meta about questions and things. And it was collecting all this data. The challenge here as a user for me is, you know, where is this going to? Is there a sinister reason for you to kind of let go of these privacies? That's what kind of brings up alarm bells for me, because I'm someone who doesn't consent to giving my private conversations in my home to you. And the fact that it's harder to opt out is a challenge. And I think that it's going to have a major blowback. And the fact that it's the 10th point, as you shared, Aaron, it's not kind of the pivotal key point that they're sharing.

It makes me not want to use their products, frankly, because I'm not 100% sure that I can trust them. And I think it's very interesting as a user, but Christian, I'm curious to understand as a developer and as you're the person and your team builds these things, how do you feel when creating the software that's essentially spying on us in our homes? Well, I mean, that brings up, yeah, brings a really good point. So, yeah, we definitely are building it.

We do use AWS a lot, but more on the infrastructure level. So like the servers, and that's where, you know, Amazon has really played a big, big role in what they have not done. And I think that they probably are trying to gear up to really fight this battle of like more like that model, AI model layer and application layer, I don't know that for sure. But when I think about, you know, the data that they are collecting, I think they're maybe trying to make a play for that. And like I said, I'm not 100% sure on that, but they need data to train on. So it's very much important that we, as consumers and developers and builders,

that we, you know, even if they're not necessarily thinking about privacy, that that is a key component. And we as developers, we make that a key component of the agentic systems that we built. So there's different components like database and the way you manage memory, but one of the components is privacy because that's not just an afterthought. It has to be a part of like the full DNA of any type of learning system. So that's something I really try to push as the genetic systems development. You know, world starts to become a lot larger with a lot more developers and a lot more companies starting to jump into the fray from the different levels, from the infrastructure level, application level.

You know, so they have to make that, you know, a key piece or people just can't trust it, right? Right. But Christian. The importance is also that you need, you know, Amazon needs it to help train the models and improve the tech and improve the services and product. But to your point, you know, it's a partnership. It's not one way street. And transparency is key here again. I'm going to keep saying that.

But when I talk about partnership, I think that there's a way. And Christian, you've probably seen this with some of your clients and some of the people that you've worked with. AWS may be an example where if it is something that's essential to their program and to their improvements, they should be offering incentivized partnerships.

So whether that be discounts, priority support, exclusive features for people that opt in, that would be a way for them to continue to retain their strong and loyal customers without them feeling like they're being used. And I think that from a business

perspective, the last thing you want is, and we've talked about this on about many different customer bases here recently, and whether it's in the airlines industry, whether it's in tech, whether, whatever, healthcare, whatever, loyalty matters to retention and growing the business long term. And so I think the partnership aspect of it is really important. What you're saying is like, you know, I need to know that that's what my data is being used for. And then I might be or comfortable to Chino, to your point, that like, okay, I want to be a part of that. Or I want to be able to pick and choose what I want to share, right? Like, you know, shopping habits.

Yeah, okay, that's why I don't care about that. But maybe there's different things. And I think this gets to, you know, even, you know, bigger issues, you know, just this week, 23 and me announced that they were going bankrupt. So it's very clear that they're selling the data. the genealogical data of all. You know, and so that's all of us or everyone who's participated. And that's a question that people, like, that's not what I signed up for, right?

I signed up to give it to you. And now I'm very concerned about where is that? I'm sorry, Aaron and I cut you. No, no. I was going to say to Christian as a developer, you know, when you put something out, especially something that has, is mass market, but depends on personalized customization, smarter and faster AI gets. Are you surprised that it took this long to remove the privacy protections because you're throttling a product that maybe could do so much more and generate, you know, perfect its generations faster if you only knew more about the customers that you were right in front of?

Are you surprised that there was so much restraint up until this point? Not really. You know, I think they're just really trying to just go really fast now. And I think they're way behind. And I think that it's more of like more desperation at this point, you know, trying to enter a category that they, I think they, that they, you know, just kind of slow rolled. One thing that they have to also really consider is that the world is actually, you know, even the U.E. And, you know, you know, everybody else besides the U.S. is really, you know, starting to actually become a lot more stricter on AI models, AI in general and just, and just data privacy and all that. So you still have to, in their global company, so you really still have to really abide by, you know, GDPR and all these other regulations.

That's just so, so, so important, you know, and that's just a regulation, the regulatory environment, you know, that's not including the public public trust, you know, which a lot of times can go hand at hand. You know, so, so it's a balancing, it's a balancing rope, you know, and, you know, I think that, you know, I think that, you know, they're, yeah, they're trying to enter a category, you know, and try to get a leg up. Because they, at the day, they're in millions of homes. And, you know, so they are collecting a lot of data, you know, and now they may see it as an opportunity to train these models. Yeah. It's interesting because, you know, Christian, you talk about the customer. And I think I'm one of the weird kind of younger millennials that I'm very, very, very. against collecting my data.

You know, AI as we know, has been known to be racist in a weird way. I'm going to say something controversial. I kind of like that because that means that you can't take my voice as a black woman and you don't know kind of the nuances of me. And that means that for my own IP, my brain, you're not just giving everything to an AI. And I think about kind of Studio Ghibli and the founder, Hale Mia Zazai, who created this. And, you know, this was back in, I think, 2016, he shared, you know, criticism of AI at the time saying, you know, this is a huge concern for the greater good.

Like, this is not a good thing, you know. And what's interesting, chat GPT, 4.0 came out with a studio Ghibli style AI that you can essentially take this entire studio studios, like years of like learning and practice and training and honing the skill. And now you can kind of cheaply recreate your own Studio Ghibli version of yourself or your dog or whatever. And I think for me, where IP is really important as my own business, as a black woman who wants to see representation in marketing, I don't want you to learn for me. And so I do, for me, with Amazon or Big Al, I'll like to call the system, I'm not going

have it in my home. And I think that's maybe a generational thing. There's some others who might say, you know what? She's a little antiquated. We need to embrace this so you can make your life better. So it's this weird balancing act as you shared, Christian, but I don't, I think there's a way in which to go about it. And I think that Amazon right now is failing because they're trying to catch up. And I don't agree with that. There needs to be privacy. If I'm not able to. to opt out or that isn't the automatic default with your systems, I'm not going to use it as a customer. Point blank. And I really appreciate you saying that. And I think that kind of goes a little bit into like, well, it does go into like really like digital identity. And and I do really feel

that that's a big part of like the promise with with AI because you can actually build, you know, have a local server into your home and and have, you know, start to actually have a learning system that's learning you and learning, like learning your home, learning your, you're learning your identity and having it all all be about you. And so I know there's a few companies starting to work on this, but I really believe kind of the next iteration of digital identity is that that you start to really start to own your own data. You know, so a lot of these companies, they are, they're taking your data as you're, you know, going about your business and in your daily life. They're taking your data using your data selling your data but but you can you know now with the help of AI you can actually

start to pull all that data down and start to start to use that generated data for your own self and and you know your own kind of autonomy and in your own agency so so I really believe that's the the near future that actually that doesn't really actually exist and it's in in the in the pure form but I believe that, and for me, even as an agentic systems developer, that's really what I want to see. That's even for me personally, I want to own my own data. Right now, none of us own our own data or know where our data is, you know, and we're generating it every day that we wake up out of bed. You know, we start generating data, you know, even while we sleep. So that's something I really, really feel, but it's super important.

So you know, you brought up a really good point as well. is that, you know, generationally, I do think there's perspective differences. I actually was just having this conversation with my children about it. And they're like, everything is getting recorded. Everything is getting is somewhere and you don't have access to it, right? So whether that's cameras on the street, whether that's, you know, all the things. So do, is there that same level of concern about privacy? I wonder about that as well.

and it's interesting because it's when I want you to use the data, not when you are using the data against me, so to speak. Because when you were talking to you know, I was just thinking of a great example that I used to use all the time when I would talk about customer experience and personalization and responsiveness. I used to travel a lot for business, pre-COVID. And so it was always on the same flight. It felt like I was always on the same, you know,

I was flying from Denver to San Francisco. every week, leaving on Monday morning, coming back on Thursday afternoon, Thursday night, you know, da-da-da-da. And what I loved about what was happening was like United knew what my schedule was. It was a pretty basic every week kind of thing. And if there was a problem with the airplane or something happened, you know, when I would call in, you'd get the automated agent, automated agent, right? And they'd say, oh, Melissa, are you calling? about your flight from SFO to Denver on Thursday afternoon. Would you like to rebook it for tomorrow morning or tomorrow afternoon? We've already got two seats on those flights because this flight

is delayed for three hours or whatever it might be. And I like that because I didn't have to do anything and I didn't have to really talk. I didn't have to really talk to anybody and they knew that I wanted an aisle seat. They knew I wanted this. They knew, you know, that even though there's a flight every 10 minutes, they knew which flights I really liked. So, you know, when it's benefiting me, I'm okay with it. But when I'm like not sure what they're doing, I think is very interesting. And so like, you know, Christian, when you're talking about that you need the data to help it learn and model things. But like, you know, like you saying, like, I don't want it to, it's not going to take who you are.

But what you're saying is like, I'd like to be in control of how my, how I, you know, present to the world or or how my data presents. Yeah. I do believe that's very true. Yeah. And I think kind of building on that, too, just to be clear, it's if you're monetizing me and I'm not getting a check from this, I don't want to give you my data. So if you're making my life better, great.

And that means, again, like, you're buying a plane ticket. You're giving money to something and it's making it easy, right? That's a service that they're providing. But if you're taking my identity, my looks, my voice, and saying, we're going to ship this off to different people than cast a check and you're not going to be involved with it. I'd rather not have my information there because I want my voice to be one words, you need to get it, right? You need to pay, you need to find someone that looks like me because not everyone looks like me and you need that as your representation. And so if that means that

you'd have to pay me to do that, I'm more open to it. But I don't like the idea of you just using me for your own money, like to cash a check. No, thank you. I think your perspective is fascinating, Chino, because that is the promise of this, you know, Big Al Plus is to learn from you, learn your habits and behaviors, and then start to mimic them and do it for your benefit without you being in the room and then showing you what was accomplished on your behalf. And all that takes, you know, to do that, takes learning from you.

I tried this a week ago, not the system we're talking about, but I tried an eight. generative system where you sit there for two minutes, you feed it some video, and you speak to it. And then I created a, we'll call it a puppet, puppet version of myself and made it say things I've never said before. And it looked pretty convincing, sounded pretty convincing. And that was a free tool. And it's, you know, early days. Now, I don't know, I signed some kind of privacy with it. You know, I clicked a box.

I don't know what they're doing with it. When I feed things to Big Al, I don't know exactly what they're doing with. with that. I know what the benefit on the surfaces that I'm receiving, but how they're, you know, what they're digging in in reselling or or data mining or all those things. Christian, you might know more than me. I'm sure you do, but I don't, I don't know what there are ulterior motives are. Well, I think so I think now we're at the point now where we can actually have more edge devices that are that are smart. So right now like Big Al like there when when you're when it is in your home, it's, going off to a server, you know, somewhere and they're doing stuff with it, right? But with, like, an edge server that's sitting in your house, it doesn't have to go to an outside server, you know, so, you know, so we're getting to the point where, you know, even companies like Amazon can basically just have a server within your house and, you know, all that data just stays within your, within your home and, you know, within your, your,

your ability to be able to train it and teach it. So I think that is an opportunity for Amazon because they're already in homes and they can make that service available. But also I think it's that opportunity for even startups, you know, where you can have these. And these are small devices, you know, just as big as big out. Just here, here you go. You know, it's there.

They can learn from you. perform actions on your behalf and really kind of have that your own kind of like digital identity be you right so so Christian isn't that more dangerous in a way to say you know even if it's even if it's offline except when you want it to right but you're still training it all the time you're feeding it data you're centralizing it you're keeping it in one place you're training you're updating it but then you allow access to it for a company that you think has you know a great service or benefit or you like what they had to offer. There's a free trial.

And then you give them full access to plug in. Isn't that more dangerous than, you know, training one app at a time and decentralizing it? It could be a, it's a single point of failure. But also, I mean, you know, you're home. So, you know, if, you know, if it's, it's on one server within your house, you know, if somebody breaks into your house, yeah, they can, they can definitely take that or something like that, you know,

but, you know, it kind of just gives you more control, you know, it's, you know, it's kind of like that it's in your castle, you know, so I think that's, that's kind of important, you know, and like I said again, like it is a, it's a, it is a single point of failure, you know, but, but it'll kind of just give you that ability to kind of know where a lot of your data is, you know, and, you know, when you leave outside of your house, you know, you're still, you still may have your iPhone watch on, that other stuff. So you're still collecting a lot of data and giving a lot of data away. But it also kind of just gives it the ability to kind of just have a little bit more, claw a little bit more data back from. Yeah. Yeah. I think that is the issue. And I think also, I'd like to know, Aaron, from a marketing perspective, because I also see, like, there's a lot of concern around, like, my pictures being used or, you know, all those kinds of things. So from a branding and marketing and advertising perspective. Like, you know, we don't want to, I don't want to share that.

But like, for some reasons, like, there are ways that people are gaining, you know, taking the photos off of, you know, Instagram or Feta or meta, you know, and using them or whatnot. And so, like, what are your thoughts in your world? Like, how do you protect that privacy, copyrights of things? you know, like it seems like it's pretty muddy down there. Yeah, it's an interesting perspective because the rules are changing in real time. I think, you know, consumers want the new shiny thing.

If it does more, especially with this generative AI, it is a generational thing. Maybe on both ends of the spectrum saying, I don't trust you, don't learn from me. But I hear it from older consumers and older adults do, right? So maybe it's those in the middle that are going to be. the early adopters, which would be an interesting, you know, dynamic shift. But now I think, you know, as long as you keep buying into it and you keep buying the latest version and you keep wanting that that personalization, you're going to get, as a consumer, you're going to get a benefit from it. For the, it comes free, the new Al Plus comes free with Amazon Prime for

now. As I mentioned, or it's a $20 a month subscription, 1999. You're probably going to get exponential value from it if you buy into it and you let it do what it's intended to do and you let it learn from you. So as a consumer, are you happy with your purchase price? Absolutely. If they stumble and misuse your data and there's a leak or any of those things, whether this company does or another one, all of a sudden, there's going to be, you know, there's a potential for a huge pushback. not just on Amazon, but on, you know, what are you doing with my data and why am I training AI models and language models and all those things? So it's, we're an experimental phase. I don't know what the answer is, but as long as these, the products and the applications that come out

have a market value that exceeds the cost, consumers are going to keep buying into it. And I think there's going to be less questions about what exactly my data is being used for. I would agree, would disagree a little bit, because, Because, you know, obviously, like, and, you know, I want to state that I'm all for AI. I do use that in my day to day, but how I want to use it. So it's very much a control thing for me where it's like, I would love it if you can, you know, capture notes for a meeting or help me, you know, draft and craft an email that I'm sending out to clients to thank them for an incredible year, et cetera, et

et cetera. Like make me more, like, you know, optimize me. I'm good with that, you know, make me a transatlantic. you know, make me a transformer. But what I'm not okay with is having something in my home. And to your point, Melissa, 23 and me, like the whole thing was, you know, we're trying to help you find your family and your genealogy. But behind the scenes, there was this larger, sinister thing going on that nobody knew

about. And I think going to your point, Christian, of having a single server, I do like that idea where I can house it. But how do I know that you're not listening? listening to things that when I don't say, hey, Bigel, you know, pipe up. I need some help here. That you're not just collecting data of my every day, all my conversations. If you have kids in the house and et cetera, et cetera, et cetera, right? So is it a matter of do we just unplug? And we plug it in when we need it because it's a robot at the end of the day. I have such a weird stance on this

because I'm pro-AI for optimizing when it comes to work and like little menial tasks in your day-to-day, like, you know, write a grocery list for me. But I don't want you to take my individual property, my likeness, my being, or that of my families, point-blank. And so where do we draw the line? And I think as a company, I do think that there's onus of them to have that ethical consideration. of sure, this can be greater good, we can teach and train these AI machine learning. But at what cost is it if you're taking away my trust, if I don't trust you anymore, I'm not going to use you no matter what the cost benefit of.

I'd rather write that grocery list. I've been doing that for how many years. So I'm good. And especially if you're monetizing that outside of my knowledge. So hopefully one day we can get into a world where, You can control it and it's confined and there's true transparency. But as it stands, that's not really the case. And with them taking back these privacy laws, it really gives me pause on using really any Amazon products, to be honest.

Well, I want to know, I worked at a financial institution around the time, this is years ago, but around the time that online banking became a thing. And there was a huge pushback and skepticism at the time about should I trust it? should I put my, you know, my banking details into your computer system? Or is this another version of that? And are we just, you know, it's just early day jitters? Or is this, you know, is this something more at stake here? Well, honestly, I think that we have to work to the future where it is, it is ours, you know, and I, and that, that is possible.

So, so I even have a Jetson Nano, which is built by Nvidia. It's a little GPU, but I can install an LLM on it and just run different scripts and just run different programs, but it doesn't have to be connected to the internet at all. And I can get the same amount of intelligence that any LLM has. And it's just on this little bitty. It's just very, very small, you know, there. And I can plug in to different programs and just other stuff, devices. And I can hook it up to a camera as well. have it see things with them in my house or whatever, but it's, but it's mine and I can give it

data and it can train. So I really feel like that that's a, that's a big, big win, you know, and I think it's also an opportunity for, for startups. Also, just even on the overall, you know, kind of identity, I think that this is the, you can actually build an agentic system that can actually go search for your, your, your image, you know, on the internet and try to claw all that, all that back. And so if you have a system that part of its job every day is to make sure that, you know, you're protected, I think that's, that's, that's beautiful. And I don't know of any system that that is doing that currently where it's your, your system, but the less that you're giving to these servers, like chat Chiquity is going to a server, you know, but you can actually do

the same thing on your, on a local device and not be sending it to chat. GPD. You know, I don't send a lot of stuff to chat GBT that I don't have to, you know, if I want to fix like a LinkedIn post, I use that, you know, but I, but it's very selective. And so, you know, so any of these systems that you're sending things to a server, there's ways to do these like locally. So I love that. And just like for our listeners, an LLM is a language learning model, question mark? Lark language. Yes, large learning model. Where I think, that idea, maybe Christian, you just came up with a brilliant new business plan, because that I would

buy into 100%, right? If I can control it, you can take my stuff down from these larger servers, because again, like, AI is the future. There's no question about that. I'm all for AI when it's beneficial. What I don't love is putting my identity, my privacy in the hands of companies like Amazon and using Big Al where they've notoriously fumbled the ball and it's been used for sinister things. And so I think you're on to something there, Christian. Thank you. I also think, Aaron, back to your question about like whether it's going to be a damaging aspect to it to Amazon.

When you look at Google, when you look at meta, when you look at OpenAI, when you look at Apple, they've all gone through these sorts of data. issues and ownership issues and they've all survived, right? So I think it's about Amazon adjusting and making sure that get ahead of the issue and making sure that it's more clear around the communications and there's more privacy controls that are allowed by their customers so that they can feel like they have some control. And then, you know, introducing like what Christian had to had to say here, you know, different types of models and different types of programs that can actually, you know, bring all of that information together in a really productive way that doesn't

feel like you're stepping on my toes and kind of invading my space without my permission, right? It's about, you know, consent. Well, let me, all right, we got the bell. Let me give it back to you, Melissa, then. we, I think we've introduced a spectrum of possibility. On the far end, we've got go ahead, do what you want to do, just provide me value as a consumer. Don't tell me about it. Then we've got the do, do what you need to do, and also keep me informed and make me a partner in what you do with my data. And, you know, keep updating me on the privacy. And then maybe somewhere in the middle

where it's, don't, you know, let me own my, I still want these features. I still want this functionality, I believe in AI, but I want to own it. And whether that's maybe something that sits, you know, in a box offline or something that you get to decide and use sparingly. And then we've got, you know, we keep going down the spectrum. Don't learn from me. You have no right to my identity. Leave me out of it. So Melissa, where do you fall? And do you think we've, we've fixed this situation that we find ourselves in? I don't know that we've necessarily fixed it. I think we've offered Amazon some really good takeaways. And so for me, I think we, the, But clearly, I think Amazon should champion privacy.

I'm not sure they will, but I think they should. And I think that they should provide clearer privacy controls that can limit the data sharing based on the user's preference. And this will help to build trust. And also, you know, you don't want to be faced with any regulatory or consumer. So that's my take of it. Chino, what's your take? So I do believe in the future of AI, but I'm a strong believer of I want to control my own data. I want to know where it's going.

There needs to be transparency. There needs to be privacy laws. For me, the end does not justify the means here. And I do think that Amazon is going to need for me to continue to be customers to make sure that that is very transparent. But I do think we're on to something in terms of allowing us or creating individual LLMs that Christian brought up earlier in the pod where you can kind of hold your own data. And I think whether Christian builds that one day or maybe Amazon looks at that and creates their own little many things, it's an opportunity. So whoever else is in the AI world, I think that's a huge market that you can tap into because there's a lot of people like me who kind of fall on both sides. Christian, what do you think? What's the best case scenario here?

Yes. So I think we're at a technology inflection point that you no longer actually need to train on people's data. They've trained enough data. So you actually don't really actually need additional people's data to really do more training. I think that's gone. Also, inference. Inference is the other part. So I think that you no longer actually really need to send it all back to the server for inference. We can do inference on that. on a device. So really, you know, the remote server, I think that that's, that's really not necessary as much anymore. So I really think that Amazon, this is a, this is a perfect opportunity, you know, to really emphasize like Melissa said, that the customer control where they say,

hey, you know, my preferences that that, you know, my data stays on my, on my device, you know, So it's learning, you know, on the device, inferences on the device. And, you know, and it stays within my, the home, within my four walls and my home. You know, so I think that's just a perfect opportunity. And then if customers want to send it to the server, they can. You know, so, but at the day, it is all about privacy and really making sure that, you know, you're adhering to that. Also, GDPR rules and all the, all the regulatory frameworks around the world.

And then also really making sure that you're being upfront and being transparent with customers, no matter who they are. And, you know, and that's really how you win. You know, so I don't think that we necessarily, you know, need in order to play in the AI game. I don't think we necessarily need to rely completely on big tech, you know. And I think that just as customers and listeners and people that are out in the world, I think that this is an opportunity to us. opportunity for us to really make sure that we start to really own our own data and own our own identity. Thank you.

Oh, thank you, Christian. I really appreciate that perspective. And thank you, thank you again for joining us. And to tell our listeners how they can find you if they want to tap into what you're doing. Yes. You know, my company's medicine analytics. We built agemptic systems for small and medium businesses. We actually recently just got our first international client.

And you can reach us at medisos.c.com. Thank you. Thank you so much. Again, Kristen Johnson's been joining us today. Thank you to Chino and Melissa. Amazon is, of course, a huge topic. So don't be surprised if we talk about the company someday again from a different angle. And that does it for another episode of We Fix It. You're welcome. Thank you for listening and growing our audience. And don't forget to tell a friend, tell a colleague, tell your voice assistant to like, rate, and subscribe to all episodes. And we will see you next time. This podcast is produced by Straightforward Media Group, All Rights Reserved. If you'd like to learn more about how a podcast

can help your company establish authority and generate leads, please email us at Eric at straightforwardmg.com or go to straightforwardmg.com for more information.