← Embracing Digital Transformation
Embracing Digital Transformation
Why AI Supercharges Collective Learning and Community Insight
00:45:41
Loading summary
Transcript162 lines
- [00:00]
A
We decided what we'd do is we'd build a platform that was sort of the anti social media so that people could have anti social with each other. Yeah, it's called deliberation IO it's all open free. You can see the science if you're interested in our code.
- [00:16]
B
So it fosters more dialogue instead of I'm going to brow beat you down type of thing. Welcome to Embracing Digital Transformation where we explore how people process policy and technology drive effective change. This is Dr. Darren, Chief Enterprise architect, educator, author and most importantly your host on this episode. Why AI supercharges collective learning and community insight with technology and social visionary, Dr. Sandy Pentland. Sandy, welcome to the show.
- [01:00]
A
Glad to be here. Looking forward to a good conversation.
- [01:04]
B
Yeah, I was excited. I went back and looked at my notes and I remember our conversation. This should be an interesting topic today. Privacy, AI, all the craziness that's out there today. But before we get started, everyone that listens to my show knows that I only have superheroes on the show. And every superhero has a background story.
- [01:30]
A
Yeah, yeah.
- [01:30]
B
So Sandy, what's your origin story? What's your background story?
- [01:35]
A
Well, I grew up in Michigan, Ann Arbor and but decided to go see the big world and have traveled around. I'm faculty at mit, helped put together the media Lab which a lot of people have heard of and now I'm at Stanford helping them with human centered AI. So AI that's good for us, not replacing us. And I think that the story is really, we got started thinking about technology and what it was going to do to people, you know, things that we're going to like wear little things like the little thing I have IOD on or know these glasses, all sorts of technology in our, on our bodies. And we got very concerned about who owns the data and what can, what's going to be done with this and is it going to make us like little androids walking around? And that's been a theme for quite a while. So, and particularly being in the context that I was in, I could see that a lot of the stories we tell ourselves about people, you know, that were rational and we work things out are just bogus. Right. We actually, most of what we do, most of what we think comes from the surrounding culture, the communities we're part of and we're, we're lost most of the time. We're trying stuff out. And so you have to think people really more as searching for answers in this context of a bunch of people who are, you know, share the same Sort of problems with you and you can, can learn from them and they can learn for you. And we're all just trying to get by. And that's not the normal attitude or story that academics like to tell. They like to tell that, you know, we're logical and so forth, but, you know, not so much.
- [03:31]
B
Well, is. Have you seen that Specifically because you've seen it probably magnified with social media. Because social media, everyone, everyone says, hey, social media is like an echo chamber, but it's an echo chamber for a reason, Right. It's the reason communities are formed. There's some commonality, right?
- [03:51]
A
That's right, yeah. No, I mean the. That echo chamber is how you share ideas with people. The title of my book is Shared Wisdom. That's how you come to a view of the world, which is your community's wisdom about the world. World doesn't mean it's true, just means it's what you all believe. But if you are in a community, like in social media nowadays, where you have these overly loud voices, people that are just screaming. Right. Like people with lots and lots of followers or people who make money off of being angry. Right. And collecting more sort of ears and eyeballs, that just destroys any sense of conversation with people. Everybody thinks about the people with the loud voices who are being paid to be angry. We've done a big experiment at Stanford where we had tens of thousands of people thinking about very, very contentious things like gun control and abortion and so forth. And if you ask people about it, they'll say, oh, those people on the other side, they're crazy. Right? But both sides believe that. In fact, both sides actually agree with each other almost completely. It's quite amazing, except for the loud guys on the edges. So when you look out, you see, oh, what are those other guys like? Well, the person you see, of course, is the one who's the loud voice, who's getting paid to behave.
- [05:21]
B
Right? Right.
- [05:22]
A
So they're all like that. Well, there are crazy people out there, but most people are like you and me. And there's some sort of. Just a little bit of sort of variation in the weight and the circumstance depends on our position in life. And so the social media amplifies the wrong things. If you actually have someplace where, you know, you don't have those overly loud voices and people can be restricted to sort of thinking about the problem, not getting distracted about 50 different things, not getting into fights, then you can have really good conversations with people. And that actually gets to the first thing I wanted to sort of Mention is we decided what we'd do is we'd build a platform that was sort of the anti social media so that people could have anti social media with each other. Yeah, it's called deliberation IO, it's all open, free. You can see the science if you're interested.
- [06:22]
B
It has code, it fosters more dialogue. Instead of I'm going to brow beat you down type of thing.
- [06:31]
A
It does two or three things. One is you can put comments in but you can't like have a reply to somebody else's comments. All the comments go in a common space. Okay, okay. And you can't tell like who put them there. And then there's a visualization of what all the comments are. So you can see where you are relative to everybody else and you can see the crazy people out at the edge and you can see generally that you're in the middle like with everybody else. And. And then we've added more recently some AI to it because all about AI.
- [07:06]
B
You have to add AI. Right?
- [07:07]
A
Gotta have to. Yeah, exactly. Right. Although we're not selling anything, we're just trying to figure out how to do this right. And what the AI does. We decided AIs, you know, make mistakes. They're, they're not really in the game, they're not human, they don't have day to day problems like we do. So we've limited it to being like a mediator or somebody who just listens and feeds things back a mirror. And so it'll say things like, I hear you people saying this, but some other people are saying that what that little bit of feedback from the AI does is it keeps people focused on the discussion and how interesting. Yeah, right. Because people are actually pretty good. This is what we do as a species. You know, way back, hundreds of thousands of years ago, we sit around the campfire talking and we talk about what did we see? What do we do? What should we do tomorrow? That's us.
- [08:13]
B
Right.
- [08:14]
A
It's just, you know, if you tried to do that with you know, a bullhorn going in your ear, it wouldn't work so well. Right. So we give that sort of around the campfire thing. And what's interesting is that like for instance, we did a town hall, doing a couple of town halls with the city of Washington D.C. and what the goal here is to hear what regular people think, not the politicians. Right. And you know, the person, you do.
- [08:42]
B
That in D.C. can you have regular.
- [08:45]
A
Are there regular people that all the politicians live in one tiny area and you can avoid Them, right? Yeah. Who knew? Any rate, so, so what we asked them is what's. What would they like to see AI doing to make government better? And you know what they answered? They answered that they would like to have an AI that they owned to fight back against all the complexity of government. I mean, keeping track of them, all that paperwork and the plans and let's do. It's like, oh yeah, I want one of those. I want something that helps me. And we call that an AI buddy. And we have a research program called loyalagents.org with Consumer Reports, which is interesting because that's a sort of thing. It's like there's all these scams out there. Who's going to protect you? Well, we're building an open source free thing that will help individuals make their own decisions and help them navigate the.
- [09:45]
B
Complexity that's out there.
- [09:47]
A
Yeah, because you know, it's crazy. I mean I'm a fairly educated guy. I look in this stuff, I work on it, I work with people. I can't do it, it's just too complicated and it changes all the time. So you know, I really sympathize with people who are, you know, have young kids and maybe they're juggling a job and stuff like that and you know, yeah, they need some help. So why, who's building something that helps them deal with the world? So that's our loyal Agents program. Well, I can talk more about that if you want. Yeah, yeah.
- [10:25]
B
I want to go back a little bit to the community thing because.
- [10:30]
A
Social.
- [10:30]
B
Media I think has just made us more disconnected in our community, but at the same time somewhat connected. But I can do it from a distance so it's safe. Where if I go down to the local pickleball courts, there's a community that's built up there. But you don't have the loud brash stuff that you have on social media because there's human contact there, right?
- [11:01]
A
Yeah. So people, people have reputations and they know if they act like. Well, I think people are afraid, you.
- [11:09]
B
Know, if I, if I elevate things, it may turn into a fistfight. Like you know, in the good old days. Right. Back in the wild west here in California.
- [11:17]
A
Right, right.
- [11:18]
B
Settled in the streets at high noon, right?
- [11:21]
A
Yep.
- [11:22]
B
So I think people, when they're in person, they, they tend to behave better. I think that's, they're in that community. Have you seen the same thing in your guys research?
- [11:33]
A
Yeah. I mean so people are quite happy to behave pleasantly. They just need to be feel like they're in a place that's safe, feel like they're in a place where they're being heard, they're recognized. And also it has to be a community. A community means people with shared interests, shared problems. Okay. It's often a physical community because, you know, the traffic in the neighborhood affects everybody in the neighborhood. There's your community. But it can also be a community of interest or a community that has particular problems. Even things like, you know, I have a rare disease or I'm in this weird situation, what do other people do? And then you're going to have conversations that actually benefit Everybody. Just like 100,000 years ago around the campfire, you know, you have conversations that helps everybody not get eaten by the lions and keeps the bellies full. And so people like that, our social media and our electronic systems in general, like government systems, purchase systems, have forgotten completely about this notion of community. You're either an individual, in which case it's, you're this little tiny person in front of the great big machine. Right, right. And in this ocean of other people that you don't know, so you, it's just sort of hard place to be, or you're all little identical robots, exactly the same. Right. So one of those two, either, you know, it's all aimed right at you, using your data to hit you as well as they can, or it's something where just everybody gets the same deal. But that's not fair. I mean, that's not about communities. Different communities need different things. A lot of what allows us to do things is having people to help us do it, which are people who think the same things, have the same problems, have the same idea about how to solve those problems. Let's collect. That's, that's how you keep things going forward. And, and we need to add that community notion back in.
- [13:48]
B
Well, I mean, when, when discussion boards first came out, bulletin boards back in the 80s 90s, it was kind of community based. Right. Because you would sign up for a, you know, a bulletin board for, you know, you know, electronic enthusiasts. Right?
- [14:09]
A
Yeah.
- [14:09]
B
For example.
- [14:10]
A
No, originally it was that way and.
- [14:13]
B
But we moved away.
- [14:14]
A
Subreddits are, are a little bit that way, but we've moved away from, because of advertising, basically. Right. Advertising as many people as possible. They don't want you to be different. They want you to all buy the same stuff.
- [14:28]
B
The same stuff.
- [14:29]
A
Yeah, yeah.
- [14:42]
B
So you were mentioning with, with these communities, now you now have an AI acting as a moderator. How, how have you seen that? Is, is that working? Well, where it's keeping the community somewhat cohesive.
- [14:59]
A
It works surprisingly well. People are respectful, they have good ideas. They, they. One of the key things is they learn what other people think because they get to see all the other comments, even if they don't agree with you. It's not like the echo chamber thing you get on most social media. And so you, you learn that there's a range of opinions and sometimes they have a good point. And you also learn that if you're going to get anything done, you can't be one of the crazy guys because you know nobody's going to follow you. And so what we see is people have these sort of discussions with this little moderation engine to sort of help summarize things, and they end up much more on the same page than they were to begin with. And perhaps even more importantly, they're much more aware of their community than they were before. So they're much less polarized. For instance, you know.
- [15:57]
B
Yeah, yeah. Has anyone brought up the fact that we've got an AI, basically not coaching, but refereeing. I don't know what the right word is.
- [16:10]
A
Moderating. Moderating. Yeah.
- [16:13]
B
Yeah. Is there any concern that you have that I've got an AI now? Because all AIs have bias, right there in some way. There's some, some bias. So is there any concern that we haven't seen it?
- [16:28]
A
Because we're not asking the AI what should we do or what are the facts. That's where you see all this bias.
- [16:36]
B
That's where you see the bias.
- [16:37]
A
Right. Because it has all this stuff, all this data built into it. Right. We're just asking it to be a mirror for what the people are saying. So you can see if it's being a good mirror or not. Because you can see the comments and you see what your AI does as summaries, and they're pretty good at that. And, and part of the reason is that they're not human. They don't have skin in the game. Right. They don't have the emotional elements to it. They don't have lots of things. They're actually not too bad as a mirror. Right. And it seems to work. In fact, when we ask people, do you prefer the human moderator or no moderator to the AI one? It's just overwhelmingly they prefer the AI one.
- [17:20]
B
I like what you said, there's no skin in the game. Because, I mean, that's one of the first kind of tenants of a moderator is, hey, I have nothing here to gain. But as a human, that's really hard.
- [17:33]
A
Right? That's right, yeah. No, everybody has anything we're talking about. You likely have skin in the game.
- [17:40]
B
Yeah, yeah, absolutely. All right, so this is really cool, a great use of AI that I've never thought of that maybe I can actually use. I chair some open, open group standard organizations and it'd be kind of cool to have an AI sit there and with no skin in the games, make sure that people are understanding each other. Because sometimes we, sometimes we need this.
- [18:06]
A
Everybody does this. It's actually. So we see people using. It's all open code. You can just use it. Right. It's not a proprietary thing. We see people using it in teaching, like, you know, the class discusses something. We see people using it in business where, you know, those endless meetings, it actually facilitates the meeting quite a bit. So, yeah, it's a pretty useful thing and it doesn't have a lot of the problems that AI have. But the other thing, remember the one that the people in Washington ds? Yeah, yeah, yeah, don't forget that one. So let me give you the sort of business version of the. Of that. They were asking for something that helped them like defend against, you know, the government. The bureaucracy. The bureaucracy. Not, not, you know, to attack, but just to fill out all those forms and stuff.
- [18:53]
B
Yeah.
- [18:54]
A
So I was on a, A, A panel with chief technical officers of some very big firms and you know, they were talking about the difficulties of deploying AI and stuff. But they. One guy just offhandedly mentioned, oh, but we've built AI buddies for everybody. And the other guy said, oh, yeah, sure. So what's an AI buddy? Well, it's just a local AI, you know, open source thing, runs on your laptop. It's not, it doesn't have to be a big thing, but it's read all the manuals that you haven't. Okay. And it reads the newsletter that you can't stand to read and it pays attention to all that sort of news. And so it can keep you in loop in your organization. It can say, well, you know, so and so is doing that over in this. Right. It can remind you about things and people are. It's very hard to be. Have all the context when you're in a big organization or even any organization. So having something that's read all the manuals and can answer your questions or remind you when you want it is actually really useful and really simple to do. And the interesting thing is if it's a local thing, you own it. Right. Or the company owns it.
- [20:08]
B
So I can put any information in there that I want.
- [20:11]
A
That's right. You can have a private thing that helps you sort of stay in the loop. And companies are beginning to understand that that's actually sort of a useful thing because you have, you know, work from home, you have offices all over the country or the world. Right. And how do you keep them all, you know, sort of engaged on the same enterprise, basically? Right. They lose context, and so they go off in different directions. And this can help them do that. Yeah.
- [20:51]
B
I'm just thinking with all the changes that organizations have, there's all this tacit knowledge that's in the organization.
- [21:01]
A
Yeah.
- [21:01]
B
That might be. It might be captured somewhere. But unless you've been with the company a long time, you don't know. Oh, the website to find out about healthcare benefits is this one over here, right.
- [21:12]
A
Or.
- [21:13]
B
Oh, time off. Oh, how do I. How do I apply for time off? You know? Oh, yeah, it's on this website over here that it's just.
- [21:21]
A
And then someone. And then the language on those websites is like, what does that mean? Am I one of those? I don't know. Is that qualified is the same thing as, like, dealing with a bureaucracy in government. It's just you need somebody to sort of be your buddy. So. AI buddies, Right?
- [21:35]
B
AI buddies. I love that. That concept.
- [21:39]
A
So the, The. If you're a technical guy or you're interested in this sort of thing, you could go to the website loyalagents.org, which is US and Consumer Reports, and then a number of other little companies that are trying to do stuff like this to basically, you know, give us a buddy to help defend against all the, the marketing and the scams and the bureau, you know, and. And there's a number of things that, you know, are important there. You have to. It has to truly represent you. So the way, say, for instance, a doctor or a lawyer word, it can't be like, you know, sort of underhandedly selling for somebody else or, you know, that's where bias can come in. Right.
- [22:23]
B
So. So that brings up. You're talking agentic a little bit. Right. They're acting on your behalf. So they become.
- [22:30]
A
These are AI agents. That's right.
- [22:32]
B
Okay. All right. So that's different than just a buddy telling you, hey, you should look at this. This is actually going to do stuff.
- [22:39]
A
This is now actually going to do things like, you know, buy a blender, fill out a form for you. Yeah. And that. And that's. That's agentic. AI, this sort of scary thing. But, you know, we're going to need it. That's like the people in Washington D.C. we're saying, God, we need something to deal with the bureaucracy, to deal with all this stuff. There's going to be all those agents out there, you know, who's going to defend you is the question. And, and the big question you put your finger on it, is how do you make sure that the thing that you're using represents you and not somebody else?
- [23:14]
B
Right.
- [23:15]
A
Okay. So that's called fiduciary duty, and that's why the name is loyal agents. The thing that differentiates a doctor or a lawyer or something like that is what's called in law, a duty of loyalty. They have to be loyal to you. And so that's a technical term in law, but it basically means, you know, they're acting on your behalf, on your.
- [23:38]
B
Behalf, for your best interest.
- [23:40]
A
For your best interest. That's right, yeah. And they're not asking for Walmart's best interest or Amazon's best interest.
- [23:46]
B
That's so hard to, that's so hard to tell now. Right. And because when you think of AI, so like if I go to a public gen AI that's out there now, I can already start seeing that they're suggesting things from advertisers. We're already starting to see it.
- [24:04]
A
So you look at ChatGPT, you can say, buy something, but the first place it goes is always Walmart.
- [24:10]
B
Yep. Yeah, I know.
- [24:11]
A
Okay.
- [24:14]
B
Yeah, so, but these are private. These are private jenny eyes that you're talking about.
- [24:19]
A
That's right, yeah, private.
- [24:21]
B
They're mine. I can, I can, I can tell that. I always like to shop at WinCo down the street.
- [24:28]
A
Nobody else needs to know about that. Right.
- [24:30]
B
And no one else needs to know.
- [24:32]
A
Stays in your little bailiwick. Right?
- [24:34]
B
Wow. I mean, this is very contrary to the way things are moving right now. Where big corporations have all my data, they advertise, they dictate what I get. This is turning the tables on them a little bit.
- [24:48]
A
That's exactly right. This is. We have an opportunity, I mean, the agentix type, everyone goes, ah, this is scary, Right. But it's an opportunity to take back some control. Because now if you have this agent that's going to sort of be trying to shop for you and do things, it doesn't have to tell those people all this data. It can just say, tell me the specs. Right. Tell me the price, tell me the this. And then it sort of thinks with your preferences privately. Right. About what will be good for you. Right. Like if you got two kids, you probably don't want a two seater car because where are you going to put the kids, right? Et cetera, those sorts of things. You don't need to share that. Right. Because your agent already has that sort of built in and it's not sharing data. So it's really a chance for user control, for privacy, for all sorts of things to come into a new sort of efficacy to be able to change this sort of new web. But it has to happen, you know, where people actually have these agents that they adopt them, that they use them. And there needs to be some sort of economic model with that because otherwise it's just going to be too easy to take Walmart's or, you know, OpenAI's model, which is going to have biases in there. And so the way I think about it, and this is why I approached Consumer Reports, is you need a band of people, a community who have shared values about what's a good thing to buy, what's a bad thing to buy, those sorts of things. And together they can trade notes about, I bought one of these blenders and it broke immediately or I bought this and this is the best thing ever. So it's a little bit like Yelp or some of those other rating things. This is a private thing among a co op, among a group of people who have banded together to sort of help them each make, you know, better decisions.
- [27:07]
B
So what you're talking about is something I've heard about or kind of surmised would exist, but you brought it to the forefront, which are community gen AI.
- [27:17]
A
That's right, that's right. This is community gen AI. Exactly right.
- [27:22]
B
And it's a closed community. You get to decide as a community who's in your community or not.
- [27:27]
A
Yeah. And if you want to be in this community or the next one over.
- [27:30]
B
Right, yeah, yeah, or the next one over. This is really interesting because I normally we think of Those community gen AIs like in medicine or in law or in a professional group. You're talking about co ops, which could be any I, I could start a co op here in my town of Folsom, California and say, hey, here's a, you know, a co op of high tech people living in Folsom, California. We start our own little co op. Our agents can talk to each other, they can share information with each other. Yeah.
- [28:05]
A
About how, how are the local stores and the local schools and all that sort of stuff where, you know, you're sharing what you feel is good for the community. You don't have to share things and you own it and you're sort of immune in that sense from, from the bias. You know, America has had co ops forever.
- [28:25]
B
Oh yeah.
- [28:26]
A
I mean, you know, it's agricultural co ops that formed the, some of the first banks that were outside of New York City and became competitors to the big banks on the East Coast. It's those same co ops later that did the electrification of America. You hear about the government doing it, but most of the electric grid was built by co ops. Right.
- [28:51]
B
That's amazing.
- [28:51]
A
Credit unions are exactly the same type of a thing, usually most of them. So. So we have a tradition of doing this and it's. It happens every time the people feel like they're being exploited by the big guys and they band together and operationalize the. The things you can do as a collective. And this is the thing you can do as collectives. You can share tips about what works and what doesn't work with people who have the same sort of circumstances as you.
- [29:25]
B
Is there any concern around security in this play? And how do I handle security in a community?
- [29:33]
A
Yeah, I mean, so again, look at the website, loyalagents.org, you know, because we talk about that. You have to have security. You have to know that this is something that's authorized by a real human, that's not some bot doing something. But those are problems that we've had for a long time. And back in the day we set up things like Open id, which is something you don't even know about. Right, May? Oh, you do? Okay, well, good. Both.
- [30:02]
B
I do know. I know about OpenID.
- [30:03]
A
Yeah, yeah. Okay. So we help set that up. And it's the thing that makes sure that your phone connects to the right numbers and you use it all the time, you're just not aware of it. And so we're working with them to extend it to these agent worlds. There's papers out there about that now. It's not that hard. We're taking the security stuff we have now, just moving it to the agentic world.
- [30:27]
B
Into digit world.
- [30:28]
A
Yeah, yeah. And then you need to know something. There's some legal parts which you need to have to say this really represents that person. But interestingly, again, we know have that legal stuff for humans. The doctor really has to do things in your best interest, not interest. And we just have to tune it up so that it works for software agents too. And so we're busy doing that with Stanford Law School and various other sort of people that are helping us. And then the final thing is, is you have to have These community. You have to have individuals or communities where you have some way of, of writing down what the preferences are. Like you don't want to waste money, you don't want to buy things that break instantly. You want to stay away from scams. All of those are preferences we have. You don't have to do that, but you'd be nuts. And people might decide to be more or less energy efficient. This is really important, carbon capture or whatever. Or they might be very concerned about kids and things that happen around and two kids. And those are preferences. And so if you have a software agent that's filling out forms or screening things so you don't get scammed and stuff, it has to know enough about you to know what it ought to do.
- [31:57]
B
And it's just been a concern that people have had. Right. Why they don't want to share because they don't want to be advertised to. They don't want a big corporation knowing so much about them. So this gives an opportunity for the agent to know all this stuff and to go out and represent you without saying who you are, per se. Right?
- [32:19]
A
That's right, yeah.
- [32:20]
B
Divulging too much information.
- [32:22]
A
Yeah. And. And there's little things in there. Like for instance, you know, sometimes websites will say, you know, we don't want any bots, so we don't want any agents. But if an agent actually legally represents you, they can't do that. It is you for all legal detentions.
- [32:41]
B
Interesting.
- [32:42]
A
So like for instance, recently Amazon sued Perplexity, I think because Perplexity's bots were looking at Amazon sites.
- [32:51]
B
Yeah, yeah.
- [32:51]
A
Their own bots. To do it. To sell you more Amazon stuff. Right. Surprise. But if you have this sort of legal determination of a fiduciary representation, then it's just you doing it. You have a little bit of software, but you have the Anyhow. Right. And so I think that will work. There might be some legal battles about it, but I think we're on the good side there. I'm gonna decide that will work. So again, putting a little bit more power back in people's hands.
- [33:29]
B
What guardrails do you have that an agent just doesn't go hog wild crazy? Because in my mind I'm like, well, should it check with me before it does anything over a threshold, if you're going to spend more than $50, you need to check with me. But if I'm doing that, then I'm. Yeah, I guess it's still I, I'm just, I want to use these bots to get work done for me that I don't want to worry about. Right. I want to focus on. On things I want to focus on, hopefully more value.
- [34:02]
A
Yeah. So there's a couple things, right, that, that you brought up. One is you want some way of rating bots. Right. Like what, what software are you going to use? Right. And sort of best practice for what is safest and most effective. So those are things that you need to sort of establish. And that's of course, what we're working on with other people. And you need to be able to have ways of communicating bots and the, the, the agents, AI agents to communicate with other agents. Because remember, they're not talking to people, they're talking to.
- [34:40]
B
Yeah.
- [34:40]
A
So there's already things like mcp, which is the way agents could talk to agents. And so we're developing hcp, which is the way that humans talk to agents. Talk to agents.
- [34:55]
B
Interesting.
- [34:55]
A
Right. So that you can say I want to buy something, but not more than 100 bucks. Right. Or I want something and here's my preferences and so forth. And that can strain what the agent can do. And it's a, you know, the simple thing is fairly simple to do the, the really complicated things where you're buying things overseas and blah, blah, blah. That gets a little more complicated. But it's no more complicated than current sort of logistics chains and things like that.
- [35:40]
B
Do you see pressure being put on e commerce sites to accommodate more and more agents? Because right now they're, they, they put up, you know, they'll put up a verified wall to prevent, you know, agents from getting in. Right. Do you see those walls breaking down? Because they may miss out. If I'm, if I'm using an AI buddy and an agent, then, and I can't hit your website, then you know what?
- [36:12]
A
Then you don't sell stuff, right.
- [36:14]
B
And you don't sell stuff.
- [36:15]
A
Right? That's right. Yeah. So, so there's this tension there. And you know, people who think about the ad economy are of the opinion that the spread of agents like this is going to get rid of a lot of the advertising because it's. The agent doesn't look at the ad. What is that about? Right. You know, the agent checks the specs and the price. Right. And the reputation with other people who.
- [36:44]
B
So that's going to be interesting because you will have marketing and you will have ads, but they'll be geared towards AI visibility more than human visibility in that case.
- [36:56]
A
Well, and remember. Yeah, but remember also the agents can check a lot of things it's like you and I get tired.
- [37:02]
B
Yeah, you're right.
- [37:04]
A
We look at three different things. Okay. And they have, like, bold claims.
- [37:07]
B
Oh, that's good enough. I'll just get that one.
- [37:09]
A
Yeah, that's right. Well, AI agents could be a lot better because they don't get tired the same way. And if you're in a co op, then you have all the eyeballs of all the people in the co op looking at these things too. So you actually get a great deal of insight about what's good. It's expected, first of all, that prices for things will tend to converge because there's going to be comparisons.
- [37:30]
B
Well, that's, that's normal economics, right?
- [37:32]
A
That's right, absolutely. It is also expected that advertising for inexpensive or sort of regular types of things may disappear because most of it will be bought by agents. On the other hand, at the very high end, like you're buying a car or house, you're going to see a very different sort of behavior where maybe your agent collects facts about houses for sale or houses that have been sold recently and, you know, arranges it in a way for you to understand, helps you understand. And that's something where it really has to be a fiduciary representing your interests, not some realtor. Right. But it's interesting because that's what realtors are supposed to do and everybody knows.
- [38:17]
B
Yeah, but they don't. Right. Yeah.
- [38:20]
A
So, so you can imagine that that will cause a bit of a revolution in that industry. Same thing with cars. I mean, you know, cars, you know, how do you really compare them? But Consumer Reports test them. Other people, car and driver test them and so forth. But a lot of that's going to end up happening through agents in the future because they're going to be able to say, well, in our co op, let's take Consumer reports, We have 4,000 people who bought Chevy, blah, blah, blah. Right. And here's the repair records, and here's.
- [38:56]
B
The, and here's all the info, which is stuff that, which is stuff that we do today. It takes a lot of our time to do all that research.
- [39:04]
A
That's right. So now you get a lot more research about things so you can make better decisions and figure out which ones are lemons or scams. Yeah, yeah, that's the hope. And, and, and, you know, it's the first time that we've had that sort of hope since, you know, sort of the beginning of the, of the web, the Internet. Right. I mean, you know, so web beginning. You know, it's like, oh gosh, you can see all these things in better choice and so forth. It very quickly became centralized and advertising driven. And so now we have another chance.
- [39:40]
B
Yeah, but I see what we can do. The advertiser guys, they're smart. They're going to figure out something.
- [39:45]
A
They make a lot of money off this. Yeah, exactly.
- [39:47]
B
Yeah, they sure do.
- [39:49]
A
Yeah. So it's, it's not clear how the battle will work, but it is clear that we better, you know, work on our side of it now.
- [39:59]
B
Yeah.
- [39:59]
A
Before they get a hold of it and completely shut it down.
- [40:03]
B
So do you think that these, A one concern that a lot of people have is that AI agents and things like this, this is going to drive even more and more isolation of individuals, you know, talking to each other and communicating with each other, because there's no need. My agent takes care of things for me now. So I'm, because I, I already see it in the young generation. If I ask my younger, my younger kids, they're, they're all in their young adults now, but if I ask them to call someone, it's like, call them?
- [40:37]
A
Yeah. What are you talking about? Someone?
- [40:41]
B
No, they, they, they want to text them. Right. Because there's a disconnect.
- [40:45]
A
Right. Do you think that this is going.
- [40:47]
B
To cause even further disconnect?
- [40:50]
A
It, it will change it, hopefully. I mean, one of the things that, for instance, in the book I talk about most is that we've done these things that just ignore the sense of community. Right. You're either an isolated individual or you're all exactly the same. So it's either federal law, everybody does, exactly, everyone does this, or you're this lonely person in the midst of nothing. And actually humans are primarily driven by community. And that's where we get our culture, that is culture. That's how we come up with innovations, things like that. And, and it's not present in most of these technologies. So, you know, it, it is something that we have to keep in mind and something to do. So for instance, the deliberation thing is intended to get people to talk to each other, to talk to each other, to tell people what to do or what to say. It is just simply a tool to help that talking together. And the AI buddy thing is, and it can be something where, you know, it's putting you in touch, it's telling you what people want to share about their experiences, people like you sharing their experiences. So it makes you more aware of what's happening. And an interesting thing is that there are some hints that this will not end up in this desolation, isolation. For instance, if you look at Facebook groups, Facebook groups almost entirely correspond to physical places.
- [42:24]
B
Yes, I've noticed that. Yeah.
- [42:26]
A
I mean, so what they're doing is they're saying, oh, all the people who are, you know, go to the basketball game on Thursday or whatever it is. Right. They're part of this group and you can talk to them by text or email or whatever, but it's a physical activity. You actually get to see the people. Reputation matters. You know, being a contributor as opposed to a troll matters.
- [42:50]
B
Right.
- [42:51]
A
And the trolls are kicked out pretty quick is what, because you can see them and, you know, because you could see. Yeah, yeah. And so. So that type of thing, people do that even in the face of a platform that is set up to ignore that people manipulate it into community. And so I think that that's why one of the reasons I like this co op idea, it doesn't matter that you're a member of one co op. You can be a member of many co ops. It's a basketball thing. Right. And it's a little co op of people that, you know, talk and try and figure out, you know, where's a good place to go.
- [43:28]
B
And the AI is there to moderate and to build, just to facilitate the conversation. Facilitate, yeah. I think this is brilliant if people want to find out more about what you've been talking about. You've mentioned your book a couple times. Is it out on Amazon?
- [43:44]
A
Yep, it's out on Amazon.
- [43:46]
B
Shared wisdom, right?
- [43:47]
A
Barnes and Noble. Yep, shared wisdom. And then there's two websites you can look at. One is deliberation IO and then loyalagents.org both of those are some of the things that I've been referring to. But the book is.
- [44:06]
B
Up. This is Sandy. This has been a great. A great discussion. It.
- [44:10]
A
Oh, thank you.
- [44:12]
B
I. I will totally be looking at these agents to do work for me, but sometimes I want to go do some of the research myself.
- [44:20]
A
Yeah, sure.
- [44:21]
B
Like, I wanna. I wanna find cheap flights. I get a thrill out of finding the cheapest flight possible. But now my agent can do better than I can, probably. So I. I got to go try that out and see what.
- [44:33]
A
Well, that's what I mean, if you go to a site like Kayak, it already does all that pretty much. Yeah.
- [44:39]
B
But. But it's got its bias in there, right?
- [44:42]
A
Yeah, for sure. Oh, for sure. It's selling stuff too. Yeah, exactly. Yeah. Yeah.
- [44:47]
B
But if I'm controlling my agent, that's even better, right?
- [44:50]
A
Yeah. Well, you could tell it don't do. Don't do X and don't do Y, right?
- [44:53]
B
Yeah, yeah, yeah. Exactly. Well, hey, Sandy, this has been wonderful. Thanks for coming on the show.
- [44:59]
A
Pleasure. Yeah. Good conversation. I've enjoyed it. So.
- [45:07]
B
Thanks for listening to Embracing Digital Transformation. If you enjoyed today's conversation, give us five stars on your favorite podcasting app or on YouTube. It really helps others discover the show. If you want to go deeper, join our exclusive community@patreon patreon.com embracingdigital where we share bonus content and you can always connect with other change makers like yourself. You can always find more resources@embracingdigital.org until next time, keep Embracing the Digital Transformation.