2025 November 11 Basic AI for seniors

You can view the original Facebook video here.

Getting Started with AI: A Practical Guide for Everyday Use

Hi, this is Jim Cranston from 7EveryMinute and 7EveryMinute.com, the podcast and website about reimagining your life. Thanks for joining me today to talk a little bit about AI and getting started with it. If you like what you hear today, please leave a like, subscribe, tell your friends, and send me a message.

Last week I was going to mention a little bit about AI and how it relates to aging and use in aging. I didn't want to mix topics up because I tend to do that quite a bit. So tonight I'm going to talk about two aspects of AI. The first is some everyday simple uses of it, and the second are some things to keep in mind when you use AI and even search results, because a lot of times the results they give look kind of correct, but they might not be quite what you were meaning to ask. So it's really anything that depends upon searching the web for information.

If you've used Google or Bing or DuckDuckGo or any other search engine, you already know how to use the basic conversational interface that most AI systems offer, such as ChatGPT, Claude Anthropic, and there are many others. Basically you type in a question and it replies, and yeah, it's really that simple.

You can always start with a free plan, but if you're going to use it regularly, I suggest getting a paid plan. Not only does it give you a little higher usage limits, a lot of times they'll give you more features such as remembering the things that you've done, so it can learn your style of writing and your style of asking questions. But the very basics of AI are very simple. It's much like asking Google something. And actually now if you look at Google when you ask it something, there's usually an AI response below it. But the difference is that most of the dedicated AI engines also offer extras that you can use, like generating pictures or diagrams or things like that—all things you can learn in the future.

However, keep in mind that if you use the memory function, that may affect the privacy settings, so you should read all that information carefully. Never, ever enter things like social security numbers or bank account numbers or any other information you wouldn't feel comfortable sharing with a stranger. Can't emphasize that too much. You probably don't want to say, "Well, I was doing some work at my bank, First National Bank on Main Street where I have an account number that ends with..." Just don't say it. If you want to say you're working with a major bank or something like that, that's fine, but don't give it anything that you wouldn't want to have somebody else see on the internet. Because if you have the privacy set to the point where it's learning from it, it can use the information you provide to give to somebody else. So keep that in mind.

Now you know the very basics of AI and you can use it as kind of an advanced search engine. As a matter of fact, there are some new browsers that have AI built in, or you can kind of use it as your own personal encyclopedia that learns the type of things that you like to look up and the way you like your answers.

But what about everything you hear about hallucinations and made-up data? Well, those stories are real, but there are ways to reduce that by how you ask questions. If you're very specific—and I'll give an example in a few minutes—and ask for references, it's less likely, though still not impossible, for you to get false information.

But it's important to remember that AI is basically kind of like a research assistant that goes out to look through a bunch of information and then gives you an answer. Now, if you asked two real human research assistants to do the same thing, they would also certainly come back almost for sure with different answers. They'd probably be similar. They would likely be two different views on the same topic unless you gave very specific guidelines as to what you were looking for.

AI engines are exactly the same. If you don't want the results to include information from before 2015, for example—perhaps you're looking at modern medical treatments—then you have to explicitly tell that to the AI system. Otherwise, it's going to guess what you meant.

Remember, one of our sayings is that the quality of your life can be measured by the quality of your questions. That applies totally to AI. If you ask kind of an open-ended question, it will make assumptions that might not be clear to you. When you do ask questions of it, you have to ask yourself why you're asking. Simple questions generate simple answers, unless that might be appropriate in some situations. But remember that AI is also capable of very complex and rich answers and information that could be a lot more helpful in really looking deeper into subjects.

This morning I was using Claude, one of the AI engines. He's the one I usually work with, and I'm going to use a query about how I formed that prompt. It's not necessarily a perfect prompt, but a lot of times if you read some of the basic AI manuals, they say, "Well, start out with something like saying assume you're a scientist in something," or something like that. If you frame the question correctly, then you don't need to take that step. It happens automatically.

I always start polite. As somebody said, does it make a difference to the results? Not so much anymore—it used to. But it's just good practice. Nice common courtesy. So: "Hi Claude. I have a lighting question related to streetlights, the non-visual cortex of the eye, and circadian activation." I told him the context of the question. And then I'll say, "I'll need any answers to have references with a link to the reference, please." A lot of the AI, particularly the older ones or some of the newer releases, if you say "I want references" and don't say that you want a link, it'll make up the reference. So it's very eager to please.

Then I say, "Assuming the streetlight has..." and give all the conditions that it's going to be operating under. The question is the details of the question. And then finally the core question. And finally, I always finish with, "Please let me know if you need any clarifications or have any questions. And thank you," because we're always polite.

You look at that and you say, "Wow, I could almost just write that myself and do that research myself." And what you'll find is he actually gave me, oh, I think it was four or five pages of information, including references and insight into some other things, and then finished up with some other items to keep in mind. That took me probably five to ten minutes to do that prompt and get it all set up. But the reality is, if I was working on my own and looking up all that information on the internet, I wouldn't be able to get that done in five or ten minutes. It saved me a tremendous amount of time getting that done. So that's one way to write a prompt.

Many of the prompts I see people writing are very general in nature. Something like, "Give me a good recipe for Christmas." And that's the extent of their question. And then they get this kind of generic answer: "Oh, well you could do this and this and this and this." Like, "Well, that was boring and it didn't even have any interesting vegetables." Well, you have to tell it that. If you're going to write a prompt, say, "Give me an interesting recipe for Christmas that includes some unique vegetables." The better you frame your question, the better results you'll get.

I saw an interesting article—and it was talking about how now I think it's 74% of the new web pages are actually being generated by AI. I have a link for it. This is a short link, but it takes you to LinkedIn. If you don't have an account, it'll ask you to join. I recommend that you do. I find LinkedIn—I use LinkedIn more than pretty much all the other social media combined. It's kind of like Facebook for adults.

But think about that: three quarters of the webpages almost are being generated by AI. It's kind of like, well, besides the fact that they're probably kind of boring and mediocre quality, the other problem is this: Remember that AI just goes through and reads all the information it can find on the web, information from books that are available, information from many sources that it can find. And if you think about all the books in the library, if you were to go in and start reading them, you probably wouldn't agree with a lot of them. Some of them may be poorly written. Some of them may just be wrong. Some of them may be providing information that you don't agree with. They may have biases that you don't agree with.

So the problem is AI has all that information. So if it writes something that's incorrect and then writes a webpage with it, later on when it goes and rereads the whole web, it sees all these webpages that have the information—incorrect information that it already put out there—and it builds upon itself that way. And you get this kind of self-reinforcing circle going on where it promotes itself. And now they're putting in some guardrails so that doesn't happen as much, so that it can tell that it's an AI-written page. But the reality is that's a moving target.

I have a kind of fun example here to show you. I did a search for Frank Medrano, and Frank is a bodybuilder and he's about 46 years old. However, Frank is also a 67-year-old actor. There are two Frank Medranos. Both of them appear on the internet, and here we have Google's set of images. These are supposedly three pictures of Frank Medrano, the actor at 67 years old. Now, if you look on the right side, Frank's looking pretty darn good for 67 years old, right?

However, when I was searching for the younger Frank's age, I found literally dozens of websites that said he was in his sixties, and it was only recently when I found a post from Frank himself—the young Frank himself—that I finally found out his correct age. But all those AIs have been training on the incorrect data. So even if you cross-checked against Google and your favorite AI, chances are the answer would be wrong.

Now I did notice very recently if you specify "Frank Medrano bodybuilder," it will give you the other Frank—the young Frank—most of the time. But it still mixes the two up with no warning. So you can see that the pictures are kind of a real visual indication that things are improving, but none of these automated systems are perfect. So always use your judgment when you're picking up any information from the internet. It's really super important.

If you think about it some more, actually one of the things is think about your favorite myth that you heard, especially that's been repeated on the internet. And you see a lot of them, obviously about boomers. It's one of the things we talk about or fighting ageism on the internet. But now again, you take any of this common knowledge that's incorrect, and again, every AI engine is reading that and it's using that for its training data. Now that wrong information is being repeated and legitimized on autopilot, kind of because the AIs are reading their own wrong answers, using that to confirm that they're correct, and you're kind of in this information spiral. It just keeps going on and on and on.

If you take this whole approach of common knowledge being repeated by the AIs, now think of it in terms of aging, ageism, elder care, geriatric medicine—all those things—and you realize that's a very dangerous situation developing.

The example that we often use is: How long have we been told that bad memory is just something that happens with age? "Of course, well, you're older, that's why you're forgetting things." Those "senior moments." And we talked about this before—that phrase wasn't even invented until 1997. It turns out the more recent research has shown that people of all ages have moments of forgetfulness.

And when people say "I'm having a senior moment," that's an example of what's called stereotype embodiment theory, and it was first proposed by Dr. Becca Levy. We've talked about her before. She's a big person on ageism and the effects of ageism on people. And what stereotype embodiment theory is, to quote: "Aging self-stereotypes form through internalization of societal stereotypes from childhood onward, influenced by family, media, and culture."

In other words, we make the reality fit the stereotype. When we were 25 and we forgot something, we just laughed about it. When we're 55 or 65 and forget something, we nervously laugh about it. "Oh, a senior moment," and think we're on the verge of self-destruction. It's a very harmful pattern of negativity and negative self-talk that does physically and mentally harm us because we're kind of fulfilling this self-made prophecy.

However, such references as "senior moments" are very common on the internet, which means if you were to ask AI about that, chances are it would agree that only older people tend to forget things.

Remember, AI's just a tool. Like every tool, it has strengths and weaknesses. If your GPS says drive down the railroad tracks, you just don't believe it. AI is no different. It can make mistakes too. Always check what you get back from AI. And remember, it can bring you a lot of interesting things and can do a lot of fun things for you—funny things and all that sort of stuff—but it's just kind of an automated personification of all the information on the internet and in books and things like that.

So you wouldn't walk into the library, open a book, go to page 104, and say everything on that page is absolutely true. You shouldn't do that with AI either. Look at the answers it gives back. Think about them. And again, remember to never tell AI something that you wouldn't tell to a stranger, maybe a casual friend, but especially if you're allowing learning from your prompts. If you allow learning from your prompts, then that information can be displayed by AI as an answer to any other person who asks it. So always be careful about what you tell AI because basically don't tell it anything that you wouldn't want it to tell somebody else.

Finally, just like people, AI can develop biases, especially since not all the information it reads is correct. So please give it a try. There's a lot of fun things you can do with it. You can get some really good information. Like I said, it did probably two or three hours of research for me in about ten minutes today. But remember to check what it says, even by taking the answer you just got from one AI and giving it to a different AI and asking it to check it. That's not a perfect way to check—bias can travel between the AIs—but at least it'll catch many errors.

And remember, there's a lot of talk about using AI in medicine or using AI to replace doctors. Use that with a lot of caution because there are a lot of biases around aging. Those have made it into the AIs as well. So if you say something like, "Oh, I'm in my sixties, I'm in my seventies, I'm in my eighties, is this normal?" that's something that—realizing a lot of things that were considered normal really aren't normal. So if you use it for medical, be extremely careful with it and use it as kind of a first pass. Look at what it says, think about it, and use it as kind of a reference before you call your doctor.

That's it for the evening. As always, thank you very much for stopping by. Your homework tonight's pretty simple. I just suggest that you work with an AI for a little bit. Try putting in some simple questions—how do you make mashed potatoes, whatever problem is in front of you. Then go on to something a little more interesting. Say, "Oh, I was planning a vacation to Europe. What's the best balance between good weather and low travel costs?" or something like that, and work up to more complex things. It takes a little while to get used to recognizing bad answers, but there are a lot of fun things you can do with it. So have fun. Give it some whirls. Don't tell it anything you don't want the whole world to know.

And of course, if you want to do extra credit on your homework, take the question you put into one AI and put it into two or three different AIs and look at how different the answers can be. You start to really get a better feel for how much they can really change one to the other.

As always, remember that one of the best ways to care for yourself is to care for others. Remember that we have UKR7.com and WCK.org. UKR7.com is for giving aid to the people of Ukraine. World Central Kitchen, WCK.org, works throughout the world helping people. Those are two great international sites. I'm sure you have local charities that can use help, but even just a smile can change someone else's day for the better in ways that you can't even imagine.

As always, thank you for stopping by. If you found something interesting or useful, please pass it along, and please subscribe and hit that like button. Please do make a comment about other things you'd like to hear. Have a great week. Remember to live the life that you dream of, because that's the path to true contentment. Love and encouragement to everyone. See you next week on 7EveryMinute and 7EveryMinute.com. Thank you.