Decoding News Credibility and the AI Revolution in Business with Chase Palmieri

Episode 44 • 22nd March 2024 • Creatives WithAI™ • Futurehand Media

00:01 - David Brown (Host)

Chase, welcome to the podcast.

00:03 - Chase Palmieri (Guest)

Hey, David, great to be here. Thank you.

00:05 - David Brown (Host)

I wanted to talk to you because, obviously, we've sort of been in contact, I guess, for about a year or so, maybe even more. It's actually, I think, longer than that when we expect. The other day, we said a year. I think it's actually been longer than that. But anyway, and I'm really curious, and I wanted to talk to you about the way that you're using AI in your tool because I think it's a really positive use of AI, and I think at the minute there's so much negative chat about how AI is going to ruin the world and how AI is terrible and all that. That. I think it would it's nice to hear somebody who's using it, I think, in a positive application. So maybe if you start off by just sort of introducing how you're using it and, you know, sort of tell everybody what you're doing and how you're doing it as sort of a foundation for the conversation, then probably a good place to start.

01:01 - Chase Palmieri (Guest)

matoes for news, and so since: 2019

01:47

Now, the way that AI comes into play here is actually what we would kind of consider step three of our master plan.

01:54

So you know, step one was to aggregate these reviews and give readers the ability to voice their feedback.

02:01

Step two was to make money by licensing out the scores to content platforms, and that's where you and I first started having our conversations.

02:10

But then step three was for us to train up an AI model based on this massive database so that we could use an AI model that could crawl any article text or any website articles in any language and actually create a stand-in credibility score ahead of any actual readers reviewing it themselves. Now this, basically, is telling us how perceived as trustworthy this particular content or this particular website will be when real reviewers end up actually reading and reviewing the content, and so it's a stand-in score for credibility that can only be achieved through the massive database of actual readers, and so that is how we try to make this AI model as unbiased as possible is by training it on data from the general public, continuous training, data still coming in every day. That is adding to the model's capabilities, and we'll dive into some of the other capabilities that our tool has, like identifying political leaning or emotional sentiment of articles and things like that. But the core functionality for credit is to identify at scale, through the human reviews and through our AI model, the credibility of content so that we can actually surface up these scores to vet the quality of inventory for brands and ad agencies who are deciding where to run their ad placements on news and information websites.

03:37 - David Brown (Host)

And the obvious question I guess that I should ask, as people are wondering, is who does? When you say people did the ratings and readers and you had journalists and professionals do the ratings, how did you select those people?

03:53 - Chase Palmieri (Guest)

Yeah, so the public score is open to anyone. Anyone listening right now could go to credder.com and pull up our news review platform and submit any article they want, or read and review any article that's already living there on the platform. So that is wide open to everyone. And there are accountability mechanisms in place that kind of you could say, force reviewers to stay on track and not get distracted and make sure that they're critically reviewing each piece of content and not just trying to attack an author or an outlet. And we have a very specific kind of a little bit of added friction in the review process there intentionally to account for that.

04:32

And then on the verify journalist side, we just look to the blue check mark journalists on Twitter. And if you have a checkmark on Twitter, that's good enough for us. If not, you can actually reach out to us and we'll manually approve you. We just need to see that you're actively writing three articles a month for any publication to date, and that will count your reviews towards the critic score. And again, we're using this rotten tomatoes type of approach where you can kind of separate at a glance see what the establishment journalism class thinks about a particular article or outlet and what the general public thinks, and that's where you see kind of big disparities every once in a while, just like you would with a Fast and Furious movie where the audience might love it, but you can expect that the critics aren't going to enjoy Fast and Furious 35 as much as the audience will.

05:20 - David Brown (Host)

Yeah, that's a good point, and I was, I guess, thinking again on the political side because you'll and, like, I've worked in or with local government and stuff for a long time, and I appreciate that you can never make anybody happy any other time. So you're going to have, you're always going to have one group saying, oh, it's biassed the other way, or vice versa. So yeah, I was just wondering how you? Because there could be some bias in there, because a lot of the media is, or a lot of people would say the media is trends left as Opposed to trending right and there aren't as many sort of right sort of leaning media. So just thinking about the bias that's in there. But what's interesting again is that you have developed a tool that can look at the actual Bias, like a bias score. How does it work?

06:12 - Chase Palmieri (Guest)

Yeah. So I would say there's two.

06:15 - David Brown (Host)

Sorry, political bias. Just to be clear, the political bias score.

06:18 - Chase Palmieri (Guest)

Yeah right, there's kind of two ways that this is accounted for. So, on one hand, in the review process, when a reader is Rating an article, not our AI, but actual readers, Excuse me they are they are Looking for any heavy partisan leaning in one direction or the other, and this is one of the ways that articles sometimes get knocked down from a what would have been a perfect five-star rating to a three-star or two and a half star, and so when a reader identifies that they're not getting kind of a balanced viewpoint or an objective take on the subject matter, then that is that that provides them the cover to leave a lower rating and that again Bakes into our AI model so that we're picking up on kind of politically jarring or Heavy-handed language within an article's content. The other thing that we've more recently added is using the GPT for integration from open AI. We've added to our toolkit the ability to Identify political lean. So what we did is we trained it up on all of the articles from within our own database.

07:26

We basically said, you know, hey, open AI we're.

07:31

We want you to be able to identify far-left, centre versus far-right content, using all of the models that it itself is trained on and using our own proprietary database of articles, and so over time, we've gotten that model to be very accurate, where now it can look at an article and not only will it give you the stand-in zero to 100% credibility score, but it'll also tell you whether it is right, leaning centre or left or actually non-political. So we have a lot of, you know, whether it's a science type of article or some other topic that is just not political at all. The AI model can identify that. Also, and the reason that we've added that is, again, we're working with advertisers, brands and ad agencies, and if they specifically want to make sure they're showing up on neutral, balanced content because of how it would reflect back on their brand, now they can target not just high quality, credible content and sources, but also the most balanced inventory out of those sites. So that's why we do what we do see in this.

08:36 - David Brown (Host)

Excuse me, this is amazing, and you know, in the context that I was thinking about it as well, and a lot of people don't know, but I've been working on this sort of social media platform in the background, and one of the things when you told me this the other day, I got really excited because it as a person who's thinking about developing a platform or a tool where people might be able to share information or whatever, being able to apply that sort of an algorithm to it, to say, actually, we don't want extreme content on either side, right, we just don't want extreme content, period, but we'll take anything in the middle, or that's sort of left to right, that's totally fine, it doesn't matter, but we don't want the extremes. So, be able to have a score that would say that Do you know what I mean? It would make it so much easier, and I think it's a lot easier, and I think it would make it a better place to be. Do you know what I mean? Just in general, and I think that's.

09:35

For me that's a proper or an excellent use of an AI tool, particularly the generative models that can do that sort of analysis very, very quickly and very accurately.

09:46 - Chase Palmieri (Guest)

Yeah, absolutely. I mean, one of the biggest problems for social media platforms, news aggregators or search engines is, as much as they can, they want to not only appear but actually be neutral and not have to make, you know, heavy-handed editorial decisions on behalf of their user base. For one, when you make editorial decisions, you actually take on liability under section 230, because you're no longer just a neutral platform, but also the backlash against, you know, using a platform that is being perceived to censor or shadow ban or hide or demote content On your behalf. It's a little bit like having kind of a nanny platform, and so what Credder allows these platforms to do is, again, Credder is not making these determinations ourselves. They're based on actual reader reviews. We're letting your own users participate in the scoring process.

10:43

So if, if a social media client of ours is using Credder's API, they can instantly Inform their content ranking algorithms to surface the higher credit credibility Scored content or source in that news feed for users without worrying about making those decisions themselves in real time or worrying about upsetting their user base, because if their user base doesn't agree that an article should be An 85 percent on Credder, that platform can simply say well, you're welcome to just go and add your reviews to that article on Credder. It's an open, transparent, third-party scoring system, and so we think that that's the the best way to make sure that we're not ourselves a censorship tool, making these decisions on behalf of, you know, free thinking people around the world, but also giving platforms the kind of at scale, instant, real-time data they need to optimise their their own platforms.

11:44 - David Brown (Host)

Sorry, hit the wrong button. Um, yeah, it's really cool. So let's take a Bigger still. Let's take a step backwards and think about the macro kind of environment. Now. I mean because you're you're in California, right, so you've probably got I mean, I'm, you know, sitting in a in a town 45 minutes outside of London and you've probably got a really different view of it, just being in in California and in sort of the Bay Area. What's the general feeling there from people about AI and how this is moving forward? Because you're obviously you know let's take a step back from you know from Credder itself. But you know you're running that company and you obviously have feelings and you know you're like this is important to you. So when you go to dinner parties and when you go out to I don't know, you know you go out for drinks or whatever and you talk to people at the bar, what's the kind of general feeling you know? Do people? Are people positive about AI? Are they apprehensive? Are they cautious? What's the general feeling out there?

12:53 - Chase Palmieri (Guest)

I think the feeling right now is that the release of Chat GPT was kind of the aha moment that the early internet had, when people were able to get free email addresses and start to communicate with each other, where the kind of we cross the chasm from the early adopters and innovators to the mainstream, realising, oh, this is here now, this is, on some level, inevitable and I have to start playing with it and paying attention to it. And I think we've seen that with companies in every sector. Now they're looking. It's like every CEO tasked somebody on their team with okay, go, find out how we can use this to reduce headcount, create more efficiencies, you know, have expand our reach, improve distribution, update performance with other aspects of our business. And so it seems like right now we're in the go taste it phase and it's not clear to me yet how all of these companies, or maybe even how the world is, is going to kind of conclude that, what their reaction to it is. But it is very clear to me that you cannot have a a kind of bury your head in the sand perspective anymore. You're going to get beaten by a young, competitive startup with two people and six AI models, and that's really the new environment that I'm seeing is that companies companies can run at incredibly low burn rates and low head counts now, and so I think the the era of like the 500 person startup or 20,000 person corporation is moving to an era of the two to five person team with an AI model and super low overhead, which allows them to charge much, much less than the big kind of established corporate competitor for maybe even a service. That is a little bit more cutting edge because it's it's a company that is based on innovation. So I think it's good for entrepreneurship, it's really good for corporate company success rate and survival rates.

15:14

You know, one thing I look at is, if I forget what the you know current number is, but say, 95% of startups fail within two years, I think that that number is going to start dropping. So if, if I only have to keep two people on payroll and an AI model and I can build unique products and services for specific niches and and maybe say my, my break even point is I only need two customers to keep my two people on payroll alive, now my company has infinite runway, my burn rate is super low and I give myself a chance to grow sustainably, maybe even without having to take on a bunch of VC capital that puts my ownership stake at risk or creates kind of these uh overt intense growth metrics and standards as as the goal. Because now, if I only have two people in an AI model, I only need to make 20 grand a month in recurring revenue and I'm putting money aside, I can keep working on R&D, nobody can take me out of business and and I can grow sustainably there and keep 100% of my company, and maybe my co-founder has 50% and I have 50%. That is a totally different makeup of how you can grow a company and keep a company alive. Now, because of these models and and I guess the obvious downside to that is, well, that 15 person startup is now only a two person startup. So what about those folks who, uh, who now aren't going to have that job? Well, what I would say and I know not everybody is in the life, you know position to become an entrepreneur, but I will say that the barrier to entry, now the threshold, is such that that 15 person startup that now is a two person startup. Well, maybe those 13 other folks become three other three person or four person AI based startups.

17:11

And when I say AI based, I mean you can work on any sector. You could be a t-shirt creation company, but use AI tools to help you more quickly, come up with designs more quickly, keep track of what the trending topic of the day is so that it's instantly you know whatever the trending Twitter topic of the day is. You instantly send that topic name to an AI model that creates a visual that represents that topic, and then that visual is instantly put on a t-shirt and within two hours, people can be buying a t-shirt that represents whatever that trending topic of the day is, and you, as the owner, are actually removed from that process because you've just built the workflow and you have these kind of models that are passing it along from step to step, and so I think it's going to revolutionise every business, but, most importantly, it's going to give the entire global economy more of an entrepreneurial base instead of kind of the top heavy corporate base that we tend to have now and for anybody listening who wants to take that idea and run with it.

18:19 - David Brown (Host)

We only charge five percent for the idea. So, uh, let us know in perpetuity.

18:28

Um, yeah, no, you're absolutely right. I think on a lot of fronts you're right, and I think one of the interesting things that you said right at the end in particular was I think this is really going to help, like a lot of the places that were entrepreneurial already, you know people who are, who have that mindset and that sort of attitude, were probably going to do it anyway. Right, whether AI was there or not, they're still going to go out and they're going to do it. That second tier of people, like you said, those 13 who now you know, kind of don't have a job to go into, um, they may be forced to do it a little bit more than maybe they wanted to, but they'll still go and probably a certain percentage of them, half of them, may go and do it anyway. And then I think what's going to happen with the other group is the other group, you know we're they're going to have to find something else to do, honestly, and I don't know what you're seeing in the US.

19:26

And again, this is an interesting thing, this is an interesting thing to talk about, but my son is so my son's high school age. He's like a seat equivalent to a senior in high school, and what a lot of the kids are doing over here is a lot of them are talking about what they're going into, like the UK has a really good apprentice program. So instead of going to university, you can sign up as an apprentice, you can go through the apprenticeship program and you still get university credit and you get a degree at the end of it. But you actually do the whole education through work experience in a company. So you actually do the job, you get paid like anybody else, but you do classes on the side.

20:13

So you do have classes that you still go to, but it's like a 50-50 kind of thing and a lot of the kids are deciding that they don't want to go to uni. You know they don't want to go down the road of doing the necessarily the entrepreneur thing themselves and they're going a lot more into trades and stuff like that. So we're getting a lot of people are going back to trade school. We're getting more electricians and more plumbers and that sort of thing, because I think you know those are the very last jobs that are going to be taken over by robotics and AI combined. Is that what's going on in the US? I mean, I don't even want to comment on what I see in the news about schools in the US, because it just looks insane, but actually like what are you seeing on the ground?

21:00 - Chase Palmieri (Guest)

Yeah, it's a really good point, and South Park had an incredible episode kind of highlighting the fact that we have all these college graduates and in the episode they were kind of highlighting the fact that they'd still go to Home Depot despite their college degree and they need somebody to actually come and fix a real thing in their house or at their place of work and that this, this kind of skill set of actually being able to manipulate, fix, treat, do maintenance on physical things in your real life, that those skills are not being taught in universities and that these degrees are only as valuable as your ability to kind of cash in on the invested capital to your education on the other side, when you're getting out of university.

21:50

And what I'm seeing, I guess and this is maybe not specific to California, but what I am seeing is that and maybe we're going a little tangential here, but bear with me- let's go I what I see is that, with inflation around the world and government debt, we are seeing older generations having to remain in the workforce for longer and longer, which means that those jobs are not opening up for the people who are coming out of college, which means people are going and spending fifty thousand, a hundred thousand hundred fifty thousand dollars on a degree, and then they're coming out and they're having to take, you know, some internship role for the first two years, or they're driving Ubers, or maybe they get into the company they want, but they're, you know a couple positions lower than what they thought they were going to be able to move into in that company when they first got out with with their fancy degree. And so the longer that the older generations have to stay in the workforce because they cannot afford to retire, the harder it is for the millennials and, you know, the younger generations to actually move into positions that will afford them the wealth that they might begin to start chasing the American dream and the idea of owning a house someday. So I think one other thing that maybe we're going to touch on is that you don't necessarily have to be an entrepreneur to benefit from this AI revolution if you're working a corporate job and you know I'm not one to promote quiet quitting or anything like that. But maybe there's going to be a new term that's like quiet AI replacement or something where you basically automate all of your typical daily or weekly workflows using an AI model or an AI assistant. And you know, as long as you show up for your weekly meeting and put your face time in, your boss doesn't necessarily know or need to know whether you are performing the work of putting together those spreadsheets or whatever it is, and there are models that will do this for you instantly and at super low cost. And so, yeah, you might have to pay out of pocket because the company's not going to pay for it, because the company's paying you to do it.

24:05

But if I pay $100 a month for an AI assistant to do the job that pays me 10 grand a month, then that's a good trade and perhaps I'm making the company more productive than we were before, because maybe they're getting the reports they need faster, maybe they're more accurate, maybe they're coming through this AI model with all these cool kind of pie charts and other things that would have taken me too long to execute on, and so it doesn't have to be a lose situation for a company, but I do think you're going to see more and more individuals try to replace much of the hours of their labor through their salaried positions with these AI models and and that that's actually a net positive for productivity worldwide and, ultimately, what AI stands to benefit humanity with is a boom in productivity that we haven't really seen since the early internet or or kind of early manufacturing and the industrial revolution, and so when you get back to this idea of kind of inflation and record debt not just household debt and credit card debt, but government debt this boom in productivity is really coming at a time when the world needs it to hopefully chew our way out of kind of this, this stale or to negative GDP growth that we're seeing across the world there's a lot to digest in that I agree, and I think what you're talking about is certainly happening in loads of places.

25:39 - David Brown (Host)

I think there's a lot of people who are using AI quietly in the background to help them with their jobs. I think the risk I guess the risk for me and part of the reason why I started this podcast in the beginning, was the problem that I see with that is there's a short-term productivity bump, but then what happens is is that the knock-on effect of that is is that, like you were saying you know these the young people are coming out of college. They've either got degrees or they don't. They're coming out of education, they're looking for jobs and those entry-level jobs they would go into don't exist anymore because they don't need to have the entry level. You don't need a junior person anymore to do pretty much anything because AI can do it. So a mid-level person who's been there a few years can use AI and they don't need to hire a junior person. So then what happens is, five years down the road, you don't have any mid-level people anymore because they never started at the beginning, but you also don't need them because you've got an AI that's now five years smarter than it was before. So now what you have is you have senior level people with AI and no mid or low level and over a 10 or 20 year period, where you end up is you end up with those companies, but you can end up with a very big organisation, or what used to be a very large organisation, with a fraction of the number of staff it may be 10% of what it used to be or 20% of what it used to be and you know, 80 or 90% of those roles now just don't exist because they don't need them anymore. And again, this is where I see that that then has a huge knock-on effect into the economy and everything else, because if those roles aren't there, then those people aren't paying taxes and you know there is a knock-on effect.

27:35

Now all of those people aren't going to be able to start their own companies because even though everybody can become an entrepreneur, like you and I, both we, you know, you've run your own company. I think you've run a couple of companies. Actually, you've had another business before. I've had my own company in the past.

27:52

It's hard, it's a grind and you know it's just because somebody can doesn't mean they'll be successful at it and I think you know the, the atmosphere sorry, there's a bug in my face, sorry the, the environment and the and the competitive environment is going to get so intense for a while, even those small startups, everybody will be using AI to create ad campaigns and to look at ads and to figure out creatives and to create images and all that. It's going to be really interesting because those soft skills, those human skills, are again going to become very, very important, because if you have 15 companies that are doing essentially the same thing because everybody can do it with AI what separates those? And then you get back to, well, it's the humans and I think it's I don't know what do you think?

28:46 - Chase Palmieri (Guest)

Okay. So on one hand, the visual of what you're kind of describing to me is kind of AI pulling up the corporate ladder. Where you can't like that ladder is getting out of reach. People can't even jump on the ladder and it keeps getting higher and higher and next thing, you know, it's just going to be the C suite with a bunch of AI tools running multi-million billion dollar companies. So I agree, the low level trying to chase a low level job to a mid level job. I think that the game of climbing the corporate ladder is over Now. Yes, knock on effects.