By: Michael Hudock

“Without tools, man is nothing. With tools, he is all.” — Thomas Carlyle

“We shape our tools, and thereafter, our tools shape us.” — Marshall McLuhan

These two quotes, by two different philosophers, from two different centuries, illustrate why this month’s blog post is so important.

Tools, as everyone know, are incredibly powerful. But the more powerful the tool, the more careful we must be with it…because all tools can be harmful if used improperly.

In my last post, I introduced a new tool that every retiree should add to their toolbox: AI. It’s one of the most exciting developments to come along in our lifetime! For retirees, it can act like your own personal assistant, making it easier to live your dream lifestyle and maintain your autonomy as you get older. But because of its power, it can also cause harm in the hands of the unwary…and be used by criminals to very bad effect.

That doesn’t mean we shouldn’t use AI. But just as you have to watch your fingers when you swing a hammer, and wear safety goggles when you operate an electric saw, it does mean we need to be cautious about how we use and interact with AI.

Last month’s post gave a number of ideas of how to use AI to enrich your life in retirement. It can help you get around town, plan and book travel, find interesting activities to try, remind you about when to pay bills or take pills, and even record your life story. But there are a number of DO’s and DON’Ts to remember as you use AI. Best practices to embrace, mistakes to avoid, and habits to build that can make AI both safer and more helpful.

DO: Be aware that AI can be prone to hallucination

AIs are infamous for occasionally “hallucinating” a response to a question by generating incorrect or even completely made-up information. (We’ll get to why this happens in a second.) It doesn’t happen very often, but it does happen often enough to make AI the digital equivalent of a somewhat flaky friend. Always available and easy to talk to, but not always the best to rely on.

Some research indicates AI can hallucinate anywhere from 3-15% of the time, which can be a problem if you’re using it to help you answer a particularly important question or handle an especially critical task. That means if you are ever using AI to help you make sense of a financial issue, draft an email, or act as your personal assistant, you should always be alert to the possibility of hallucinations.

Along similar lines, remember this:

DON’T: Automatically trust the recommendations AI gives you.

Because AI is so fast and so helpful, it’s extremely convenient to turn to when you need a quick answer to a question or are looking for advice. But when you do this, don’t automatically assume the answer you get is the right answer for you. Here’s why.

When we talk about AI, what we’re really referring to is something called a Large Language Model, or LLM. Without getting too deep into the weeds, an LLM is a type of mathematical model designed to function similar to the way a human brain does. To do that, it is trained on a vast archive of text scraped from the internet.

Whenever you interact with AI, the LLM behind it uses all that text to calculate the statistically most likely next word in a sentence. Essentially, it’s a form of pattern-matching, like the autocomplete function on your phone when you write a text, except much, much smarter. (This is why AI occasionally hallucinates. If it lacks the correct data to answer a question or follow a prompt properly, it will “guess” instead based on the data it does have.)

This technology is amazingly powerful, and incredibly useful! However, because everything AI generates is based on statistical probabilities, it means the answers or advice it gives is not specific to you. It’s not based on your specific needs, situation, or emotion. (Although the more information you give to an AI, the closer to you its approximations will be). AI doesn’t know you. It’s just really good at statistics.

That matters, because true financial advice must be personal in order to be good. It must take into account who you are, and who you are is unique.

So, while you can certainly use AI to find information, brainstorm ideas, or get recommendations, don’t automatically trust that it’s a fit for you. Here are some examples, based on some common financial questions people ask AI:

“What is a Roth IRA, and should I have one?”

AI can certainly tell you what a Roth IRA is, but its response on whether you should or should not have one may not be correct for your situation.

“How can I reach my financial goals faster?”

AI can definitely give you some compelling ideas here, but its suggestions may be all wrong for your specific financial situation. For instance, if AI counsels you to put more money into the markets each month, that might help you reach your goals faster…but it could also lead you to take on more risk than you can personally afford.

“Is it time to put more money into the markets?”

There’s no one-size-fits-all answer to this question. Now, a good LLM would probably acknowledge that, but as a result, it still couldn’t give an answer that is truly specific to whether it’s time for you to put more money into the markets.

Because of all this, here’s what you should do:

DO: Use AI to check your thinking.

“I’m thinking of starting a new workout regimen. What are the pros and cons of joining a gym versus working out on my own?”

“I’ve heard that the cheapest day to fly to Europe is on Wednesday. Is that cheaper or is there more I should know?”

“I love the way hydrangeas look, but I’m concerned the climate might be too dry for them. Is it still possible to grow them in a shady spot in my garden, or should I consider something else that looks similar but needs less water?”

This kind of approach is far better. Here, you are not asking AI to tell you what to think. You are asking AI for a second opinion on what you’re already thinking about, giving you more information to make an informed decision. In this case, AI is a resource, similar to a reference book, a website, or an app on your phone. It’s a way of helping you create a plan rather than creating a generic plan for you.

DO: Use AI to organize, filter, and find the data you need.

Imagine you are planning a trip and have all sorts of data to keep track of. Dates, flight times, check-in times, costs, what’s been paid, what hasn’t, when things are due, etc. It can be almost impossible to keep track of it all in your head, and extremely time consuming to organize it all yourself.

This is where AI excels.

When it comes to finance, organizing, filtering, and finding data is one of the best ways to use AI. You can upload all that travel information, and AI can organize it into a nifty spreadsheet for you in just a few seconds. You can scan a document and ask for a high-level summary, or to pull out specific items from the text. You can give it a logbook or financial report and ask it to find all instances of a particular entry or data point. For all intents and purposes, the limits are defined only by your own needs and imagination. AI is an incredible time-saver, coordinator, and organizer!

DON’T: Share personal information like SSNs, PINS, and other sensitive data.

With all that said, never share overly-personal information with AI. Anything you upload, enter, or even type with an AI chatbot is stored on the parent company’s servers. That means it can be exposed in data breaches, accessed by third parties, or used as part of the LLM’s training data. This is especially true for Social Security numbers, PINs, passwords, medical history, lab results, and proprietary business information. For this reason, be especially careful that any documents or messages you share with AI do not contain this sort of information. Even blacking it out is no guarantee of safety!

As a general rule, treat AI the way you do the rest of the internet: Be cautious, protect your privacy, and don’t overshare!

DON’T: Fall for AI scams.

Have you ever seen a picture on your phone and wondered if it was real or just AI? What about a video with a celebrity in it?

AI is great at creating things that seem real, but aren’t. The most sophisticated of these are known as deepfakes — synthetic videos, images, and voices that seem incredibly real but aren’t. For most people, this technology is a harmless toy. But for criminals, it can be an incredibly insidious tool.

There are many ways that crooks can use AI to defraud their victims, but two to be especially watchful for are personalized phishing scams and AI voice-cloning. With the former, cybercriminals scrape the internet for your data, then use LLMs to craft emails and text messages that perfectly mimic the stores you shop at, the financial institutions you use, or the people you work with. All it takes is one convincing message that seems to be legitimate to get victims to unwittingly give up passwords or even wire money.

Even more sinister is when con artists take videos off social media, feed it into AI, and then create recordings that sound almost exactly like a friend, boss, or loved one going through an emergency. When you pick up the phone, you’ll hear their voice say they’ve been kidnapped, or they’re in prison, or they missed their flight in a foreign country. “Could you please wire me some money immediately?” the voice will say. Or, “I need to get into our account but I can’t remember our password, what is it?”

For these reasons, everyone — especially retirees and pre-retirees — need to be more watchful than ever for cybercrime. Whenever you see anything on any of your devices, especially a message, video, or image that seems to compel you to take some sort of action, always stop and ask yourself: “Is this AI?”

***

AI is a powerful tool, one with so many possible uses. It can ease and enrich our lives, especially in retirement. But, like all tools, it can also be dangerous when used improperly.

Here at Hudock Capital, we use every tool at our disposal to help our clients plan for the future they want. After all, without tools, we are nothing! But our tools can also shape our perceptions, experiences, and outcomes in ways we may not intend. That’s why, when it comes to your finances, tools like AI should be used as part of a larger, more comprehensive strategy. A strategy that’s personalized to you, based on your exact needs, resources, and goals. A strategy steered and safeguarded by a team you know and trust.

We are honored to serve as that team for many families in the Williamsport area. So, if you would ever like a second opinion on your strategy, or even on the information and insights you’ve gleaned from AI, please think of us. We would love to meet with you!