Lately, I’ve been thinking a lot about search engines. I remember using Alta Vista and Yahoo!, and I remember my first time using Google. It’s like the internet flooded the world with information, and search engines became to the internet what the Dewey Decimal System is to libraries. That’s the natural evolution – first the collection of information, next the organization of information, then eventually the curation of that information. AI seems to be the next natural step of curation in the maturing life of the search engines.
From OpenAI’s Chat GPT to Google’s Gemini to Microsoft’s Copilot, we are seeing heavy adoption, and these AI solutions are becoming more and more mainstream.
Here’s what concerns me. When I was simply using a search engine, I was more of a skeptic/critic. I’d plug in a search, study the source providing the information, and determine whether the information was credible. It was this second-nature due diligence that I never even thought much about. I was always a curious skeptic, whether I was searching for travel tips or restaurant recommendations. Why? Because Yelp had Del Taco as one of the top-rated restaurants in our city. I just learned to always verify.
For some reason, I’m not sure I can fully explain it, but I find myself more trusting of the auto-generated AI answers to my search queries. Rather than a footnoted URL, I’m just getting a short-form answer to my questions, and I kind of blindly assign a higher credibility to those quick-generated answers. Let me be clear, I absolutely should not be more trusting, which is the primary motivation for today’s article.
Today I want to share two short stories from my personal life. From these accounts, I want to draw attention to the importance of always verifying before trusting.
First Story
So, I’ll have to start out by embarrassingly admitting that my family and I are total Disney fanatics. We love the movies, the park, and the magical nostalgia of it all.
Some friends of ours have been kind enough to invite us as guests to an exclusive club within the Disneyland Park. A club that is said to have a 10-year waiting list, reserved for Disney royalty alone.
The club is elegant, magical, memorable, and simply encapsulates all that is Disney. If you are getting the vibe that my family would love to be members, you are spot on.
We’ve been referred by members, applied online, and I’ve impatiently followed up to see if I could potentially nudge my way up the waiting list.
Last week I wanted to follow up and I was looking through my sent messages to recall the club’s email address. I couldn’t find it at first glance, so I ran a quick search on Google, and sure enough, the AI bot generated the email address quickly for me. So, I sent over a check-in note and was elated by the quick reply that presented me with the opportunity to join. I exchanged a few emails, and then I found it quite fishy when they were asking me to wire membership dues to some obscure foreign account.
I was being duped.
To calm your worries, I did not send a penny as I realized this email address was fraudulent. I was fooled, at least for a moment, because the website name seemed legitimate, the actual website showed photos and a credible description of the club, and the real kicker was that I just blindly trusted this AI tool as a shortcut to obtaining the email address.
Following this, I went back through my emails to gather the real email address with which I had shared exchanges previously. I informed the folks at Disney about the situation and encouraged them to take action.
Second Story
Last week, my colleague Blaine Carver wrote a wonderful piece on Marriage and Money. On the podcast, Blaine mentioned that Travis Kelce and Taylor Swift’s recent engagement announcement had garnered 30 million likes on Instagram. I don’t personally have any social media accounts, so I curiously asked if that was the most “likes” of all time, and if not, what was? Blaine didn’t know the answer, and I was still curious, so off to the search engines I went.
In my first search, Google’s “AI Overview” told me that a photo of Lionel Messi celebrating Argentina’s 2022 FIFA World Cup win was the most liked post in Instagram history. I scrolled a bit further and there was another common query, “What is the most likes on Instagram in 24 hours?” So I clicked the expand icon, and this is the answer that was generated:
Source: Google
I was absolutely laughing out loud when I read this. Again, I know very little about pop culture and even less about social media, but I surely knew that was not the right answer. Yet, I could see the tangled web that the AI bot was weaving.
Yes, it’s an incredibly funny error, but one that’s inconsequential.
When Money is on the Line
The humor fades away when these errors and half-truths lead to financial loss.
With the rise of AI adoption, I have started to see folks leaning on these tools for financial advice. Some questions are as simple as when to claim Social Security, and other questions are more nuanced around the complexity of estate laws.
I ask you this: if an AI bot thought Travis Kelce proposed to Lionel Messi, is it fair to say that there could be these same stumbles around investing or financial planning questions? The answer here should be a resounding yes.
Does this mean that we throw out the baby with the bathwater? No, there is a place for these tools, but we must be careful where we place our trust; we must always verify. These are tools to be used, and we must wield them wisely.
A Bunch of Bull
I was sharing a few of these stories above with someone in the AI space, and I like the way he summed it up. He said that at this point, based on where technology stands today, AI is a professional bull_ _ _ _ ter.
I laughed at this description, but it is the perfect description. The scary thing about dealing with a BS’er is that, often, their claims sound believable, and they use anecdotes and vocabulary that can really draw you in.
It’s like a professional poker player. A majority of the time, they will wager because they have a hand that warrants their bet, but sometimes they are just bluffing. If they bluff all the time, they become the little boy who cried wolf and just lose all credibility. It’s those sprinkled in bluffs that will often fool their opponents. Sure, AI isn’t your opponent; it’s not seeking to intentionally dupe you, but you can’t always tell when it is bluffing. You can’t always see when AI is a bit over its skis, which means you need to be REALLY careful on how much you rely on these tools.
Good Advice Looks Like This…
To wrap up today’s article, I’d like to take a quick look at what good advice actually looks like.
It was fun to tell these stories and laugh a bit about them, but the key here was to remind all of us to be careful where we place our faith.
When I think about “good advice,” here are the qualities that come to mind for me:
- The advice giver has an education and experience in the field associated with your inquiry
- The advice giver has “skin in the game,” whether that is financial, reputational, social, etc.
- You are able to hold the advice-giver accountable
- The advice can be reasonably understood; not mastered, but generally comprehensible
This list is not exhaustive, and I am sure on the podcast, we will add a few other thoughts, but this is a good starting point when assessing where you are gleaning your advice from.
It’s a crazy world out there – fraudsters, half-truths, lions, tigers, and bears – proceed with caution, my friends.
