← Lead With AI
Lead With AI
The War on Misinformation: How AmICredible Is Becoming the World’s Fact Filter
The War on Misinformation: How AmICredible Is Becoming the World’s Fact Filter
Loading summary
Transcript70 lines
- [00:01]
A
On today's episode of Lead with AI, it's all about something we desperately need in the digital age. Trust. With misinformation spreading faster than ever, what if there was an AI that could tell you before you hit share, whether your post is actually credible and better yet, whether or not a post you've seen is credible. That's exactly what my guest, Dan Nottingham and his team have built with amicredible, an AI platform designed to fact check your content, rate your sources, and help you build digital trust one post at a time. Because in a world where everyone's got a platform, credibility might just be your most valuable currency. Let's get into it.
- [00:51]
B
Welcome to lead with AI. I'm Dr. Tamara Nall. In each episode, we we will take you behind the scenes with visionary leaders shaping the future of AI across public and private sectors. Join us as we explore groundbreaking projects and innovations that are transforming industries and making a real impact on people's lives. Let's dive in.
- [01:22]
A
So, hi everyone. Welcome back. It's Dr. T, your host with lead with AI and we're so excited because number one, we actually hit the number one spot in technology on Apple podcast this year and we have won the W3 gold award in interviewing for podcasts. So we're so excited and it's listeners like you that make it possible for us now to be award winning as well as number one ranked. So I'm also excited because we couldn't get here if we didn't have wonderful guests like our guest today. And I'm excited about talking about am I credible? But we have Dan Nottingham, who is the founder and CEO of micredible. Dan, how are you?
- [02:08]
C
I'm just fine, thanks. And thank you for having me on this award winning program.
- [02:14]
A
Thank you, thank you. So tell us a little bit before we get into the product itself. Tell us a little bit about who you are at your core, your passions, and how your passion kind of led to founding am I credible? And knowing that there was a need for it.
- [02:34]
C
So if we go back to my core, we have to go all the way back to my college days where I majored in physics and astronomy and I got a job at Boston University center for Space Physics, where I worked for seven years as a staff scientist. And so at my core, I am a scientist, although my career took me away from academia. But during, you know, with this scientific mindset, I became really frustrated hearing people say things like, I don't know what to believe anymore. And I understand that. And when you take into account that trust in mainstream media is at an all time low. People dismiss expertise in favor for some opinion based belief. That that really sort of frustrated me. And so am I credible? Is what I created to sort of push back against the wave of misinformation and mistrust. Just my way of trying to do something right?
- [03:36]
A
No. That's amazing and it's interesting. I actually, one of my alma maters where I got my doctorate is George Washington University and we had a breakfast alumni event this morning and we were just talking about how it is so hard to know what's true and what's credible. And there are all these different data sources that are coming in and statements that are being made. And even with AI you don't know if it's real or not. Which depending on who you are could be a good thing because you can just blame AI and say that's not me, that's AI. But going to, trying to figure out what's real. And so I, I am just getting goosebumps, I'm hearing about am I credible? So tell us a little bit more about what it is and what it does.
- [04:20]
C
Well, it tries to mimic what a regular person would do if they were really trying to research something to find out whether or not it's true. And so what you would do if you were good at doing research, you would start with your sources and you would check credible sources. So what MI credible does at first is, goes out and it uses some technical approach to determine what's a credible source for this particular piece of information you're asking about. And so it'll reach out to those multiple different sources. And it purposely is designed to look for varying points of view. It brings that together because we're trying to weed out bias the best we can, brings together all that information into one place and, and then has multiple large language models pass over that information to do an analysis. Is the statement that you're trying to have validated or the question you have, do all these sources agree? Is there consensus that you can build? And of course the large language models, you might argue, well, some of them have bias in them. The way they're trained, the way they prompt it. So that's why we use multiple large language models to then kind of look through the data and decide what is. What am I really learning from this? And then at the end it creates a full analysis of what it's learned. It lists all the sources and creates a credibility score from 0 to 10 that tells you how credible is this? So at a glance, before you Even read the analysis, you can see right away is what I'm researching, what I'm trying to say, what I'm thinking about, is it credible? And then you read about why or why not.
- [06:07]
A
Okay, well, you're just showing you're just sharing all the secrets. Because I was going to actually ask that later in terms of if we were to open up the hood and look at the brain, how it works. And so you share that with us. And let me just make sure that I'm clear. So if I want to pose some fact or even some opinion on social media or even talk about it and try to sing cool at a cocktail party, I can put that in there and say, you know, one out of ten of every, you know, American or everyone has X, Y and Z and I can put it in am I credible? And it would actually tell me whether or not what I'm saying is truth or truthful.
- [06:46]
C
That. That is spot on. You got it. So the basic idea is one, you can do what most people seem to naturally want to do is check other people. Right? Yeah. See if they're saying the right things. That doesn't sound right to me. Let me check that. You absolutely can do that with am I credible? But one of. I think that the, the use that I'd love to see come out of this is that people pause for a second before they post, proactively deciding, I'm going to make a statement. I think I'm right. Let me just double check before I post it, before I.
- [07:23]
A
Right.
- [07:23]
C
I might. How do I know I'm not spreading misinformation?
- [07:26]
A
Correct.
- [07:27]
C
And if you have, if you're going to try to be a credible person, you know, put the. Be accountable for yourself and what you say, maybe you'll check things before you say it. And that way you kind of proactively can stop misinformation from right in its tracks before it starts spreading, before it's even created.
- [07:50]
A
Right, Right. No, I love that. And then, too, and we were talking about this before the show, is that I can then take other people's posts and their statements and check it as well.
- [08:02]
C
Yes.
- [08:03]
A
And actually that probably will lead to some interesting debates and conversations because I can then. I can then say, well, actually that is not true what you posted and hear the sources. And then I don't know why, but I'm. I'm actually a little bit obsessed with being, you know, being kind of like cool, like cocktail parties and stuff. So not only can I say the truthful, credible fact, I can then say well, according to source A, B and C, you know, so that's good. No, I, and this is so timely, Dan. Again, given everything that's coming out in social media, I, I sometimes just feel so overwhelmed, you know, and I feel like sometimes, particularly now, people are drawn to where they're already aligned. Right. And so this would, this is a great AI platform to be able to check myself and have some self reflection on kind of like my views and my thoughts and whether or not they're accurate.
- [09:04]
C
You know, you, you make an interesting point in the cocktail party where you, you, you hear someone say something and you want to challenge them but maybe you're not confident enough your own knowledge that you, you kind of let it go or, or that even happens online where you see something. Oh yeah, I don't think that's right. But I'm not confident enough to challenge it and I know everybody's going to just jump on me so I'll just stay away. The idea here is that maybe we can start having more civil conversations by bringing them down to fact based things, you know, instead of challenging somebody and say, oh you're just wrong. And let me tell you why you can say that's interesting. Here's what am I credible has to say about your statement. You know, I didn't do it. But here's right as I see them now you're starting to bring in some facts into the conversation and you can have a more civil discourse other than, you know, just start yelling at each other or not saying anything at all.
- [10:07]
A
Yes, absolutely. Though I'm not one to not say anything but yes, I should move towards that. But yeah, now what platforms because so is it an app on my phone? So can I literally be at a cocktail party and put something in there or do a voice command and get.
- [10:22]
C
The facts in real time, yes or no. So yes, you can do that, but no, it's not a mobile app yet. It's a browser based so you can bring up your browser and do it on your phone but Perfect.
- [10:37]
A
Okay, great.
- [10:38]
C
We're building the app now so it'll be available on Android and Apple soon in the next few months, but right now users can bring it on and I use it all the time on my phone. It works fine. Just going to be better when we have the app.
- [10:55]
A
Awesome. I love it. Yep, so we can do that there. So let's now talk about the holy smokes moment. Tell us about a time where a user or a tester or someone actually used Am I credible? And it changed everything. For, for them.
- [11:09]
C
So I'm going to give you an example of myself.
- [11:13]
A
Okay.
- [11:14]
C
Because, you know, I'm not eavesdropping on people using it, the kinds of things they're asking and stuff.
- [11:19]
A
Well, I didn't know if you received some email or some communications like, wow, you know, those.
- [11:24]
C
Not yet. You know, we released in September and I have heard that I personally got to experience what I hoped the application was going to do for others. And I was with a group of friends and politics came up and this group, we're very diverse politically. It's kind of fun. We all have different views and we all get along well. I was having a discussion with one of my friends who, we were polar opposites on one particular thing, very common these days, doesn't matter what it was, but we didn't agree on what had actually happened. And so I, after a little back and forth, put it into am I credible? And turned out I was wrong.
- [12:14]
A
Okay.
- [12:14]
C
It was actually correct. I had misinterpreted the event and I shared that with him, that, yeah, hey, you know, I, I was off. I learned something. And then we're able to continue on the discussion, but the discussion became quite different now. I had learned something and I could actually delve into it more deeply, understand why I didn't understand what really happened, how I had misled. And it was a kind of a growth experience for me. And I'm hoping that's the experience that others will have using this.
- [12:50]
A
Right, right. People can, like, come, come down and be humble about it. Okay, that's great. Now, through the development of Am I Credible, were there moments where you're, I mean, you gave us a wonderful example there where you were having this discussion with your colleague, your friend. During the development of Am I Credible, were there moments where you were like, wow, I can't believe that I founded this, that I developed this. Like, this really is a game changer. And can you just take us back to that moment and if, you know, you have a victory song or dance, feel free to do that as well. You know, however you celebrate, you can share us. But yeah, take us to that moment where through the development of the product, you were yourself amazed and at. At. At what mi credible can do.
- [13:37]
C
At first I thought this was going to be something that was going to be very easy, that, you know, you just prompt a large language model properly and it's going to come back with and it's just saving people the trouble of creating prompt. Right. So that there's not a big technical hurdle here. And I tried using Chat, GPGPT and Gemini and a few others and thought, okay, you know, if I just prompt it correctly, it'll give me back some answers. But then it turns out when you really start hammering it with all kinds of different sorts of information that you might want to get out of it, we discover that it's just not that simple. AI will kind of go off and it'll hallucinate it'll. Things wrong. It'll, it'll. It doesn't behave quite the way you want it to. So it required almost a year's worth of work of, of understanding how the engines work in this particular context. Yes. AI is an emerging technology, so it's not always going to behave the way you think it's going to behave. So we had to do a lot of different techniques. A lot of proprietary code was written to the AI engines behave in a way that was going to give us the results we wanted. What the big moment for me was when I was putting in just silly things just to check it, just simple things like the earth is flat or something like that, and see what back with. After a great deal of effort to kind of tune it so that it would actually behave consistently the way we want it to, the analysis started coming back very deep. And the depth of the information that it was feeding back was more than just a something quick that you might get on a prompted version of ChatGPT. It was much deeper. It was citing sources like in one case, it was citing Fox News and MSNBC in the same analysis. Okay, you could see it trying to weed out biases. You could see it actually coming out with credibility scores that made sense. That, and the exciting moment to me was that I learned, just as I did with that example with that friend, that we had created something that isn't. Wasn't any of our own worldviews, all of us working on it, we're discovering that none of our opinions were baked into this at all. We were all learning from it. And I remember that the first time it came back with the full analysis with a credibility score that looked like, just felt right. It was spot on. Now we have something. It doesn't matter what the rest of the product looks like. That analysis and that credibility score with its source material, that is the product, that's the information that we're trying to share. And that was exciting when it all came together and it worked.
- [16:50]
A
Got it. No, that is amazing. Now, if I put something in the MI credible and I get my credibility score, what's the range of what's considered credible versus not. Is it 8 to 10? Is it up to me as a user? How does that work?
- [17:04]
C
Yeah, so it's a spectrum, right? It's from 0 to 10, 10 being the most credible thing and 0 being. I can't find anything credible about this statement at all. The facts are wrong, the context is wrong, and it's different from fact checking. You know, you could, you could put in. And this is where misinformation is, gets so tricky sometimes you can list several facts, but then you can wrap it in a context that distorts it. So you can say all these facts are correct, but the context in which it's being used is misleading. And so we use all of that to create what we call credibility score rather than just fact checking. So it's things together and we have a. It's kind of a proprietary algorithm to determine the score itself that ranks it. But if it's unsure, if it says, yeah, this is a little iffy, that's going to be somewhere in the middle of the scale between a four and a six. Things over six tend to feel like, okay, this is more credible than not.
- [18:13]
A
So I'm trying to wrap my hands around there is a difference, I hear you, between fact checking and credibility. I get that. I'm just trying to find the difference in my mind and reconcile it. I guess if you could have three facts and all of them are incorrect, but if your context is trying to point them out that they're incorrect, then that would be a high credibility score. I'm just do kind of. See, I'm trying to.
- [18:47]
C
So the facts and the context have to both be correct for it to come out with a high credibility score.
- [18:57]
A
And what's an example of a context?
- [18:59]
C
Okay, so let's say here's a fact. I walk outside and I see the horizon spreading out before me and that the. The sky kind of meets the horizon way off in the distance, looks flat. I look overhead, looks like there's a dome. I have two facts that everyone agree on. The the plane ahead of me looks flat and the sky looks like a dome, therefore the earth is flat. Both of the things I said were facts and my conclusion. If those are the only facts, that conclusion would seem credible if it weren't for a whole breadth of other facts that were ignored. Okay, well, I wrapped a couple of facts with incorrect conclusion because chose to ignore a bunch of other facts.
- [19:55]
A
Got it. Okay, that's very helpful. Thank you for that. Now, how do you think about ethics? What are the ethical Considerations that you've considered while, you know, developing and innovating. Am I credible?
- [20:09]
C
Bias has been our biggest challenge. Okay, got it. So bias and transparency are two areas that we really focus on. So bias is the hardest one, because who. Who's making the call into what's biased and what's not, what's true? You know, you don't want to pretend you're the oracle of truth. There isn't such a thing. Truth has context. Truth is kind of a slippery thing. So how do you. That's why we chose to use the word credible.
- [20:39]
A
Credible, yeah.
- [20:41]
C
So trying to weed out that bias was just a way of being able to say, let's just do what any reporter would do, any good reporter would do, with some general journalistic integrity. They look at multiple sources, they try to check with multiple different tools. They look across all kinds of different points of view, and then come back with a conclusion. Now, if you were to do that as an individual, that would take you the better part of a day to do that one particular thing. So we wanted to try to embrace that idea of trying to not stack the deck. Don't give it a worldview. Don't try to make it lean one way or the other. Just find out what's true, what is. Well, I shouldn't say true, what's credible, what isn't. And by averaging out all the biases as much as we could, we believe that was a way of being as ethical as we could with this, not to support any particular worldview.
- [21:45]
A
Got it. And does. Am I credible? Does. You talked about prompting earlier. Does it depend on how I prompt or is just so focused on giving it a credibility score that I just put in the statement and it automatically knows what it's doing. I don't have to, like, prompt it.
- [22:01]
C
That's it. So we do have a prompt that we built into it, but it says things like, the prompt includes statements like, you are. You do not take a political stance on anything. You are. To be an unbiased judge, you are. You know, we try to tell it what it is and what it's trying to do. So we've. We've built all that into it to try to be as neutral as we can. All you have to do is put in a statement and it evaluates it.
- [22:28]
A
Got it? Okay. Awesome. Now, where do you see the big future for mi? Credible? You know, five years from now, three years from now, a year from now? What is the big future? And how is micredible going to change the world?
- [22:40]
C
What we see as the future is a place where credibility and trust is kind of built into the system. So right now you think of social media platforms. Information spreads through interactions with it, right? Positive or negative, we want to see or we see being added in the future, that credibility, something with a high credibility and high interaction spreads something with maybe a lot of interactions. But a low credibility is, doesn't spread as much is froth. Now that doesn't mean that we tell you what to think or that we try to stifle free speech. It's just a matter of being like a responsible editor. You know, there are things that aren't worthy of being published. And yet all of us now with social media are kind of global publishers, right. And we have nothing giving us any guidance on what you should be publishing. I see credibility in the future being used as a tool for being able to tell something's interesting, something's popular and credible that spreads more easily than something that is interesting, popular and not credible.
- [24:06]
A
Got it. Okay. Fair. Fair like that. And if our listeners want to try out am I credible today? How can I best do that?
- [24:17]
C
Well, you simply go to amicredible AI okay. And right there you can go straight in as a guest and try it out. I think the best way to judge am I credible is to use it. Just see what, what it says. But you don't even have to create an account to try it. But you like what it does. You can then create an account and then you can start to build your own credibility score based on than things that you check and then share.
- [24:45]
A
Oh, okay.
- [24:46]
C
If you don't share it, it doesn't count against your credibility score. But things that you share, you can start building up a credibility score of your own. And there's also a pro version that the reason there's a small fee associated with it is because it costs money to go as deep. It's a much deeper check, deeper search. So you go to the pro version, it'll do a much deeper analysis and it will also allow you to to compare yourself to other users as a leaderboard to see credibility scores and usage rankings in globally or by region. Anyway, it's just kind of fun for just engagement.
- [25:24]
A
Right, Got it. And you were talking about you can, you can use it as a guest and you can build your own credibility score if you share it. Now how does it know whether or not I shared what I my statements or my facts or.
- [25:39]
C
So you can share from right within the tool. Okay, create an account. You can go in there because now it knows who you are. You can go in there and you can say, okay, I've made a statement, boy, I want to tell people about this. You press share and it'll give you a list of different social media apps that you can. It'll just post it for you.
- [25:55]
A
Got it, Got it. Okay. I love that. And then we always have a question from one genius to another. Dan, you are a genius for this week and our previous genius has a question for you. And that question is, when you think about AI in your industry, what keeps you up at night?
- [26:14]
C
My main concern right now is its misuse. It's a tool like any other. You can use it for good or for evil. Right now people associate AI with misinformation and spreading it, you can certainly use it for that. But it can also be used for good to fight misinformation, such as with am I credible?
- [26:35]
A
Got it. Yeah, no, that's definitely. I can't truly say that I lose sleep, but that would be something if I did that, that I would be concerned about. So let's move to our bonus rapid fire. I'll give you four questions. I'll quickly list them off to you. Give us the first thing that comes to mind, starting with what is the most overrated AI or tech trend?
- [26:56]
C
I think it's every product must be AI enabled to remain relevant. Ah, I don't think that's some AI isn't the right solution for everything?
- [27:07]
A
Yes. Okay, got it. Awesome. What is the most underrated AI tech trend?
- [27:13]
C
The rise of small, specialized AI models. Large language models can do an awful lot, but it's those specialized models that go deep instead of broad.
- [27:24]
A
Okay, awesome. What is a book we should read about the future or AI or just a good book?
- [27:31]
C
Well, you know, I like the Righteous Mind by Jonathan Haidt. You know, it explores why people believe different versions of the truth and how understanding that is key to overcoming misinformation.
- [27:44]
A
I love that and I love all your answers tie back to am I credible? So that's good too. Well, I haven't asked the last question yet, but yes, so far. So what is your biggest, boldest prediction?
- [27:56]
C
Prediction. So this doesn't necessarily tie back to am I credible?
- [28:00]
A
Okay.
- [28:00]
C
But I, I believe that most software in the future will be written by AI. Humans will just describe the intent of the software, but really most people won't be writing code.
- [28:13]
A
Awesome, Awesome. I love that and I love am I credible? I'm definitely going to go sign up after this because I have a lot of opinions and, and what I think are facts. And so this would help me with my credibility score and the credibility score and seeing what other people are posting. So you told us to go if we wanted to go try it out. That's amicredible. AI. How do we. What are your social media tags? Both you personally and for Am I credible? And are you on LinkedIn? How can we engage with you?
- [28:46]
C
So my Dan Nottingham AM on LinkedIn and and Facebook and what have you and so is am I credible? So okay. It has all the same pages.
- [28:58]
A
Awesome, awesome, awesome, awesome. Well, Dan, I have so enjoyed this conversation and learning more about am I Credible? And honestly, it is a very timely AI product given all of the things that were going on, not just here in the US but globally. And people are overwhelmed and pretty much burdened about what they find to be truth versus not. And so it is very clear to me that MI Credible is really changing that for everyone and is a highly needed AI product. So thank you so much for being a guest here on Lead with AI and it's been a wonderful discussion.
- [29:38]
C
Thank you for having me.
- [29:39]
A
Absolutely. All right, everyone. So until next time, Lead with AI. Bye.
- [29:46]
B
Thanks for tuning in to Lead with AI. I'll see you next time as we continue exploring the cutting edge innovations shaping AI across the public and private sectors. Until then, keep Leading with AI.