Avoid AI lies and save your credibility with these must-know strategies for entrepreneurs! In this video, we dive into practical ways to protect your reputation while leveraging powerful AI tools like ChatGPT, Perplexity, and Claude.
If you’re ready to transform your content strategy, supercharge your business, and take your entrepreneurial journey to the next level, you’re in the right place.
Discover the truth about AI limitations, learn how to combat AI hallucinations, and explore simple yet effective methods to verify information.
We share actionable tips like multi-source verification, consulting subject matter experts, and using AI tools for entrepreneurs to ensure accuracy in your work.
Whether you’re crafting proposals, creating content, or solving business challenges, these strategies will help you maintain trust and credibility.
Ready to elevate your entrepreneurial strategies and embrace business growth tools that work? Join the conversation and subscribe for more tips! Let’s continue building smarter, stronger businesses together while staying ahead in the ever-changing world of AI in business.
Stay tuned for more insights, and let’s tackle the future with confidence!
#news #dwnews #bbcnews #usnews #communicationsdecencyact
Recommended AI tools for business growth.
http://AItools4biz.com
Here’s a link to the video:
https://youtu.be/mKYAJsg5VIY
Listen to this & other episodes on our podcast
https://bit.ly/sre_podcast
CHAPTERS:
00:00 – Intro
01:03 – AI Chatbots Limitations: “I Don’t Know”
06:01 – AI Fact-Checking: Practical Guide
10:38 – Information Verification Techniques
12:52 – Iterative Verification Process
18:14 – Importance of Peer Review
21:25 – Staying Informed on AI
21:53 – Building Your Credibility
22:45 – Closing Thoughts
For our friends, amigos, who enjoy the Spanish language (I’m loving leaning it!) here’s a summary of this episode in Spanish:
En este episodio de “Stark Raving Entrepreneurs”, titulado “Avoid AI Hallucinations”, Terry Brock, el anfitrión, discute cómo evitar las “alucinaciones” de la inteligencia artificial cuando se utilizan modelos de lenguaje como ChatGPT en la creación de contenido. Terry comparte la importancia de verificar la información antes de confiar en ella, ya que basarse en datos incorrectos puede dañar la reputación profesional.
Resalta la necesidad de aplicar principios de periodismo al utilizar herramientas de IA, como verificar con múltiples fuentes y distinguir entre las confiables y las no confiables. Terry sugiere usar una regla de tres fuentes independientes para corroborar hechos. Además, menciona la importancia de la humildad en la IA, sugiriendo que admitir “no saber” algo puede aumentar la confianza en estas herramientas.
El episodio aborda cómo mejorar la verificación y asegura la credibilidad de la información utilizando herramientas como Snopes y plataformas comunitarias para debatir y mantenerse actualizado sobre las mejoras de la IA. También se recomienda fomentar la revisión por pares y participar en comunidades que apoyen el aprendizaje constante. Terry enfatiza que mantener la credibilidad requiere un proceso robusto y continuo de verificación.
And for your convenience, here’s a transcript of this episode for your learning and enjoyment as well.
Terry Brock [00:00:00]:
Okay. Picture this. You’re crafting a really important client proposal. You’re using chat GPT, and you look at what it says about something, and it comes up with a study. I’m gonna use air quotes on that. Study that says such and such happened. And then we find out later, woah, wait a minute. That doesn’t exist.
Terry Brock [00:00:19]:
Oh my goodness. Suddenly, your reputation, poof, up in smoke because you relied on information that was not true. Well, I’m gonna show you in this video how you can turn that around and turn that liability into something really strong for you by doing the right things, embracing some of the principles of journalism that I learned in school that can help you. Hi. I’m Terry Brock, and we work with entrepreneurs how to solve problems like this and many others. In this video, I’m gonna show you some specific steps that are really important if you’re using ChatGPT, Perplexity, Claude, all kinds of wonderful tools that are out there, those large language models. But you gotta be careful that you don’t run into the problems. Otherwise, your hard earned credibility, out the door.
Terry Brock [00:01:04]:
There was an article recently in the Wall Street Journal that I wanna share with you that has some really good perspectives on this. Let’s take a look at that article. That article was in the Wall Street Journal, and I thought it was pretty nice. Let me pull this up here on the screen so you can take a look at it and see what it is. It’s it was entitled, why do AI chatbots have such a hard time admitting I don’t know? He had someone. Well, maybe it’s because I like so many people. I don’t know. We’re afraid that if we say I don’t know, it makes us look bad.
Terry Brock [00:01:32]:
Actually, it enhances your reputation when you can say, I don’t know about certain things. I think that’s important. We’re gonna get into that more later. He talks about hallucinations here are the hottest problem in artificial intelligence, spurring companies and research to find new solutions. I thought he raised some very good points here. As we get into the article and look at what he’s talking about here, a lot of times people will wonder who is your who are you married to? Ben Fritz is the author of this. And Ben, by the way, if you’re ever watching this, thank you very much. Good job on it, buddy.
Terry Brock [00:02:03]:
Really appreciate what you and the Wall Street Journal folks are doing. But it’s a simple question. However, he finds that it’s, really convoluted and twisted around. I think he said it really well here when he went into the idea of despite their ability to solve some of the world’s most complex math problems and convincingly simulate human relationships, AI chatbots regularly get facts wrong. They invent legal cases, jumble the facts, famous movies and books, and, yes, they make up spouses. So all kinds of problems can come from that. You might think, well, what are we gonna do? We thought it was accurate. We’ve done Google searches on that, but, things have been usually good.
Terry Brock [00:02:42]:
But there’s also some problems with that too. It’s kinda like the old joke that we have. Said, oh, I got it on the Internet, so it must be true. Well, no. Hey, scooter. Sometimes it isn’t that way. You know, it can be out there on the Internet and it’s not exactly what it says it is, but there are ways that you can cut through that. We’re gonna take the latest in technology of AI, couple it with the sound principles of journalism, putting those together to give you a way to handle this when things go wrong.
Terry Brock [00:03:12]:
And when they don’t quite go the way that we want them to, and we see that, okay, there’s some real problems here. We’ve gotta turn around, and we gotta make sure that, we do it right. Well, I think he said something really good. We’ve been doing that in that article when he put it in there. But, I wanna make sure you understand you can get some help you need on this and other areas over at starcravingentrepreneurs.com. Bounce over there, take a peek at us, and see what we’ve got because we can help you on these kind of issues and many other areas also that you’re working with. So we’re gonna get a chance to look at that. But let’s take a look here at what he said toward the end of the article.
Terry Brock [00:03:43]:
He said, insofar as you can take that knowledge about its own limitations and try to get it to convey that it looks like the best solution, Amanda Askell, an anthropic saver who helps train Claude’s personality. I think this is really good. Claude, by the way, is a great little tool for this. Even as AI has gotten more powerful, Americans’ faith in it has been decreasing. Now here’s something I would question. He said in the article, in 2023, ’50 ‘2 percent of people were more concerned than excited about AI, and that’s according to Pew Research Center. However, I would quickly come back and say, I get it, but, hey, 02/2023 is a long time ago in AI history. Hey, anything even last week is almost too old.
Terry Brock [00:04:27]:
It moves fast. So we need up to date information, and that’s one of the things we’re gonna talk about there as well. You see it. However, I think there’s a big part of the solution here, what he said, and I have to agree with this article very, very much here. When he says, giving AI more modesty could be part of the solution. When someone asks you a difficult question and they say, I cannot answer, I think that’s really good. I think when you could say, hey. I don’t know the answer to that one.
Terry Brock [00:04:54]:
That’s perfectly alright. I think that builds trust, said professor Hernandez Orallo as he observed. We are not following the common sense advice when we build AI. What we’ve gotta do is build that in there. We’ve gotta learn to say, hey. I don’t know. For instance, I live in Orlando, Florida, and I’m aware of what’s going on here in the state of Florida, etcetera, like you probably are where you live. But if somebody said, well, Terry, what do you think about such and such really important issue on the ballot right now in Boise, Idaho, or in Albuquerque, New Mexico, or in Petoskey, Michigan, wherever it might be that we’re looking, frankly, I don’t know.
Terry Brock [00:05:33]:
I don’t keep up with all that stuff. And saying I don’t know shows that you’ve got some respect. You’re not gonna say, well, I think they ought to do this, or I think you ought to vote for this candidate or that one without doing your research. Not good. One of the ways you can really earn trust, as the professor said, is you wanna make sure that you’re following the evidence. And many times, just saying, frankly, I don’t know. And that’s okay. It’s a good answer.
Terry Brock [00:06:00]:
Well, there are several things we can look at on that. What we wanna do is take a look at some of those here with you, and then AI fact checking to look at a practical AI for this. I put this together based on what I’ve seen, got the information and help from chat GPT and from perplexity to put all this together for you. And I’m using gamma, by the way, my wonderful new heartthrob of screen sharing. So or not screen sharing, but putting together presentations. It does a real good job of that. And I wanna go through this and show you some of the things that you can be aware of and you can do look at. Because what we’re gonna do is we’re gonna explore the best practices for content creators and consumers to combat those AI hallucinations.
Terry Brock [00:06:41]:
You’re gonna learn to verify information effectively. This kind of thing is important. First of all, we gotta understand what in the world is AI limitation. What does that mean? Well, we’ve got hallucinations, and these are AI models that generate incorrect information. Yep. They do that. You’d think, okay. They’re not gonna do that.
Terry Brock [00:06:58]:
It’s gonna give me the right thing. Not always. Often, it will give you something that might not be quite correct. And so what you’ve gotta do is you’ve gotta find a source that’s reliable. Reliability. The LLMs, those large language models, struggle to distinguish reliable and unreliable sources. Okay. Hey.
Terry Brock [00:07:18]:
They’re machines. Hey. I know people that are struggling on that too. And some of them, they let vote, but that’s a whole another thing right now. But, they didn’t ask me if they could. Okay? What you wanna do is you wanna do a little rule that can help you out real fast right away here, and that is the idea of looking at multi source verification, the three source rule. What that means is you want independent resources. You wanna look at people that are separate, not tied to each other, that they corroborate facts with three reputable sources.
Terry Brock [00:07:48]:
You wanna do that. Reputable sources, and they don’t know each other. This is what they told us in journalism. They wanna make sure you have three. Doesn’t mean all three are gonna be right. All three could be wrong. But what you wanna do is you wanna increase your knowledge base and reliability. Now if we take that using AI and go way beyond just three, that’s the beauty of it.
Terry Brock [00:08:09]:
Today, in using AI, you don’t have to go to just one or two or three sources. Use a variety and use multiple LLMs, different sources. So when you do that, then you’re gonna be able to get more reliable information all the way around. I think that’s gonna be important. And secondly, look at distinct origins. Hey, take a look at where it came from. Ensure that the sources are non corroborating, that they’re not getting involved with each other and going, hey. I’ll tell them this and you say the same thing.
Terry Brock [00:08:37]:
Okay? Make sure there are three sources completely separate from each other. And consult your primary sources as well. You wanna do some original research on that, looking at trace information to original research papers. That’s something you can do very well with tools we’ve talked about here at Start Craving Entrepreneurs like DeepSeek, that one that is based in China, but you can get servers in The US, Canada, and Europe so you keep your privacy. I just had to throw that in as an oh, by the way. Really important. But you can get that information. Also, take a look at official documents.
Terry Brock [00:09:11]:
Use government websites to find out, okay, what are they saying about this? What do they say here? Does the FDA say that? What does the IRS say about this? What does it say about copyright? There’s a copyright.gov that my buddy, Russ Riddle, told me about. He’s a lawyer that works in copyright law, and that’s been the help. Russ, thank you very much for watching this. You can go over there, find out what the government is really saying about it. But go to those documents, go to those websites to see what’s going on. And then other authoritative institutions that’ll give you the information you need from time to time. Oops. I missed that one over there.
Terry Brock [00:09:43]:
Let me come back here. Hey. It’s just us here. Use as benchmarks. Those authoritative institutions, use them as benchmarks so that you can say, okay, this is kinda where we ought to be and what we ought to pay attention to. And then request the source references. Citations, please. Where did you hear that? That is one of the things I love, love, love about perplexity.
Terry Brock [00:10:05]:
Perplexity gives us the ability to see where they got the information. Where did this one come from? Where did that come from? We can find out that information very quickly and easily. It just shows you the source. Now the source could be wrong. It could be inaccurate on a given point or more, but at least you know where it came from. So that makes it much better to find out what’s really, really going on there. And always ask for supporting citations. When you’re looking into what needs to be done, check on where it came from and, why we should believe that.
Terry Brock [00:10:37]:
And then cross check your references. This is important. Got three important things here I wanna mention to you that you wanna pay attention to. Number one, publication date. Take a look. When did this thing come out? Did it come out just yesterday? Credibility as well as context. With a publication date, verify. Did it come out like I just showed you with the Wall Street Journal two years ago? It could still be accurate.
Terry Brock [00:11:01]:
I don’t know, but I’d like to see something a little more up to date than about two years ago as we’re doing this recording. So we wanna find out what’s going on there. On the credibility, you wanna assess the credibility of that source. Is it coming from someone that is known to be credible? And every time you’ve heard her talk, she was right on target. I mean, she was just incredible. She did that. Okay. That’s a reasonable reason to believe that it’s true.
Terry Brock [00:11:29]:
But other times, it’s not quite gonna be there. So you wanna work with that. And then on the context, ensure that it’s relevant. Not just, hey, we’re talking about something over here that could be true, but if it’s not relevant to the case, not good. For instance, I know a lot about AI and marketing. That’s what I talk about. But if you need help on heart surgery, uh-uh. Not me.
Terry Brock [00:11:51]:
I did not we did not study a whole lot of heart surgery when I went to business school. Okay? They left that part out. I’m gonna trust certified real medical professionals who specialize in heart surgery and heart heart attacks, and they know about all that stuff. These are the people you talk to. So make sure that you’re looking at what is going to be real and what’s gonna help you in the real world getting the right kind of information that will help you in many, many different ways. Also, consult subject matter experts. Find out people who, again, really know this stuff. Get the expert advice that they have and confirm the accuracy for your specialized topics.
Terry Brock [00:12:31]:
Just because someone says they know it or they read it on Google or they read it on chat GPT or Perplexity or Grok or wherever it is, doesn’t necessarily mean that it’s right. You need to do your homework. You need to find out, okay, is this something that I can believe? Can I really test this and make sure it’s okay? This is gonna help you to be a lot more credible and keep that reputation of yours in good shape. Also, you want iterative verification. This is important when you work with others. Working with someone to say, what did you get? Oh, I got this. Really? I got something a little different. Let’s look at that here.
Terry Brock [00:13:04]:
Oh, and what did we get over here? Working in groups can be very helpful with that. What you wanna do is, first of all, take a look at what is going on in the areas of working with different kinds of cross reference. Find out what’s happening there. What is being said in many different areas? What one party says and what another one says. And then use different sources like chat GPT, Perplexity, Grock, Claude, and there are many other wonderful sources out there. There’s quite a few and it can get overwhelming. But I recommend that you have those available in your hip pocket. One of the ways I do it, unashamedly, I am a customer of Perplexity.
Terry Brock [00:13:43]:
They’re not paying me to say this. I like what they’re doing to give you access to the different tools. So you can go over here and go over there and over here too to do some comparisons. Don’t take it, as truth when you hear it only in one place. Also, be your own fact checker. Fact check what’s going on. Look at things like politifact, Snopes, and others. Snopes, I think, is one of the, really outstanding people and groups out there that can check to see, okay, this is what’s going on.
Terry Brock [00:14:15]:
This seems to be true or this doesn’t seem to be true. Check that out and have those kind of sources in your hip pocket so you can check on them quickly. Also, take a look at your source records. You wanna keep records of the verification steps. You wanna make sure that what they’re doing as they go through it is credible. And provide disclaimers. Look at your AI source content. This is important.
Terry Brock [00:14:37]:
Hey. When you use AI, let people know that you’re you you use that. What I did is matter of fact, for this, I checked with chat GPT and perplexity, brought in some of the best of it, then I ran it again through gamma to make sure it’s working well. Then I look at it myself and edit it again. So I look at what the AI gives me as a rough draft. It’s the rough draft and then I go in and analyze it and check it. That’s what I do on my newsletters that I send out every Tuesday, Thursday, and Saturday. Lots of good information.
Terry Brock [00:15:08]:
Thank you very much for the kind comments I’ve been getting lately about the kind of work we’re doing there. Thank you. But those are available and we wanna make sure that we check the sources so we can go, okay, this is what’s happening. This is where we need to be and where we shouldn’t be. Look at the sources on that and clarify AI’s role in the verification process. Let people know when you have used AI. And so you can say, this is from AI. Fair enough.
Terry Brock [00:15:34]:
We understand that. But try to say this is from me when you got it always all of it from AI. Not cool, Sparky. You don’t wanna do that right there. And then here’s a real important one. Really important. You wanna make sure that you have critical evaluation and skepticism. Yep.
Terry Brock [00:15:52]:
Be kind of skeptical. I want you to kind of notice, does it feel right? Is this right or does there see something that seems like is not quite right with that? I gotta say this is something that’s important. I think I could be speaking painting here with a broad brush, but I think women tend to have a little bit better insight into what’s going on than men do many times. Not always, but quite often. And women say, okay. What we’re gonna do here, is this the correct answer? Doesn’t feel right. What I would recommend is you take both. Take your gut feeling, which can be wrong, but you get that, and do your research, which can also be wrong.
Terry Brock [00:16:27]:
Put the two together and then have collaboration with others, and you’re gonna do a whole lot better all the way around when you start working with that. You’re gonna find that it’s much better. And then assess the plausibility within the existing knowledge that you have. What kind of information are we getting from that? Is it good and can we trust it? And then check for bias or for outdated data. Sometimes, it’s just way past the time. It’s old. It’s no longer true. You need to verify that.
Terry Brock [00:16:57]:
And that means you’re gonna use a lot of common sense. Just because you see it on the Internet, well, it might have been from the Internet ten years ago and not true anymore. Do your homework. Find out what is gonna be true and what’s true today. Check for your bias and notice the context because that context really matters. And consider the time sensitivity. Are they saying this is hot news now, but it’s from a 1936 newspaper? Well, hang on. You know, that’s not quite gonna be right.
Terry Brock [00:17:26]:
And then use multiple AI tools. For instance, you might say, wow, I use tool a and it told me this. That’s a pretty good answer. Yeah. Tool b came in just a little bit better. But then there’s tool c and, I wasn’t quite as good as the others. Okay? So weigh them back and forth. I’ve done that on this channel here several times when I look at different ideas.
Terry Brock [00:17:48]:
Here’s one way you can do it. Here’s another. It’s okay. But not in this case, they might be not quite as accurate. But this one, hey, it looks pretty good. Part of that journalism thinking is gonna pay off for you. So we take the best of AI and the best of tried and true principles that work. First principles and as I have mentioned, some of the journalism, real journalism that really works.
Terry Brock [00:18:11]:
You make sure you get that together. And then here’s something also really important. Encourage peer review. Get others to take a look at it. Something we talk a lot about here at Star Craving Entrepreneurs where we say the old Japanese saying, none of us is as smart as all of us. You wanna make sure that when you’re looking for something that you don’t rely only on yourself. As smart as you might be or one person as smart as she might be or he might be. You wanna look at what’s going on in many different places.
Terry Brock [00:18:46]:
Get a variety of different opinions. Find out what they’re saying and use that. None of us is as smart as all of us. Very, very important on this and we wanna see how to get that. One One of the ways you can do that is oh, hold on. This is here. Yeah. Get a second or third or fourth or even more opinions.
Terry Brock [00:19:06]:
Sometimes that’s a good way to do it. You wanna find out what people are saying. Are we getting the same thing from all these different places or is this guy the only one that’s saying that? Now he might be right. Sometimes that happens. But often if this one person is saying it and everybody else is saying something different, not always, but usually that person could be wrong. Then you wanna collaborate with your colleagues. This is what excuse me. What we’re talking about here again, review view your information, see what’s there, get feedback on the sources, find out what’s going on in many different ways but how you can handle that and what can be done.
Terry Brock [00:19:41]:
Then use community platforms. I love doing this. Get online feedback. One of the ways you can find out what people are really thinking is to go over to places like Reddit, x, formerly Twitter and over to, Facebook and other places. Find out what people are saying what they really believe. Often what they’ll say in one place isn’t exactly what they think. They’re putting in what they think is the right thing to say. Oh, is this what I’m supposed to say? Whereas you get over in other places like Reddit and maybe substack in many areas, you’re gonna find out what they really believe.
Terry Brock [00:20:14]:
So be sure and check on other places you can go and that online feedback will give you really good information in many ways. Then engage in those discussions. Find out what people are saying, why they’re saying it, and get information from them, giving valid information to them, giving help that way. And then keep updated with AI improvements. As things change, can’t keep hating on to what was good last year. I know last year seems like pretty relevant, but it could be out of date in the tools of AI. It’s one of the things we do in our community of star craving entrepreneurs. We keep things updated.
Terry Brock [00:20:50]:
Every Wednesday, we have a tantalizing tech tip from Terry, me, where I’m saying, here’s something new going on. Here’s something new that you can use as an entrepreneur. Not just here’s something flashy and new. Those are neat and groovy, and, the geek in me likes to hear those. But realistically, how can this help you build your business? That’s the key. And so you want to keep up with those AI improvements in a way that’s gonna keep you relevant to what’s going on in a very strong way. Follow the updates and stay informed. Real important to do that.
Terry Brock [00:21:26]:
And as you’re monitoring the updates, here’s how you can. Look for new features that are out there. Scan for market relevant evidence of what’s out there. And then for improvements, they might the tool that you looked at last year and thought, yeah, it’s just not good. They might have brought in new management or brought in some new people or they heard that you and others didn’t like it and they’ve now improved that. So learn from the changes that are there. Discover what’s being done and how you can make it even better. And then ensure your credibility.
Terry Brock [00:21:56]:
This is gonna be important. You wanna mitigate the risk of relying on AI hallucinations. Make sure you can put those aside. And then build a robust process for regularly checking this to see what’s going on. And then, maintain your credibility unlike what we said at the beginning of this video that you can ruin your credibility by trusting something that’s not true. You’re gonna maintain your credibility by doing it in the right way. So our AI fact checking guide in a practical way, you need to explore best practices for content creators and consumers and combat AI hallucinations. Learn to verify that information effectively.
Terry Brock [00:22:38]:
What you wanna do is make sure that you’re staying in touch with what’s going on. And there’s ways to do that. One of the ways would be to get involved with those communities we’re talking about. So if you look at closing thoughts, this is important. Hang in there with me. This is gonna be real important as you look at what is available. Keep abreast with what’s going on. Get involved with the community.
Terry Brock [00:22:58]:
Gina and I offer stark raving entrepreneurs where we’re learning a lot just from the people that are there. But also, we find that we’re learning as we tap into our own communities to learn other new information. Find a community you can get into and learn from it and keep learning constantly. That’s gonna be the way to do it. Stark raving entrepreneurs, we’re here to help you all grounded in the live and let live principle. Don’t hurt others. Don’t take their stuff. Do not initiate force or coercion.
Terry Brock [00:23:27]:
Live peacefully. If you embrace peace, alright. Where for you there? You get out of sign of peace, I don’t care for that. Rather be peaceful. So do that. And by the way, speaking of growing and learning more, got a video here that’s gonna help you a lot. This is gonna be something that’s gonna show you how you can tap into other areas of AI and start working with it in a big way. I’m Terry Brock.
Terry Brock [00:23:50]:
Thanks for joining me here for Stargraving Entrepreneurs, and we’ll look forward to hearing from you.