On Valentine’s Day my brother-in-law kindly wrote a poem to his wife, and then other poems to several others of us in a particular chat thread. At least one person was moved to tears by the tenderness of the poetry. Another, my other brother-in-law in the conversation, rolled his eyes and lauded his brother’s ability to pass off content created by ChatGPT as his own.
It was tender, all around, because “you’re an idiot” is how these two brothers say “I love you” to each other.
Unless you’ve been under a rock, you’ve probably read something about tools like ChatGPT and it’s ilk, such as Microsoft’s new Bing and Google’s Bard.
I have a few takeaways that I’d like to share on the topic.
1. Things are about to get weird.
Yeah, that’s a real takeaway. Just like it was strange for my brother-in-law to be writing poems, there are going to be other weird things going on. I predict that we’re going to see a lot of bizarre stuff happening. Part of this is that it will become difficult to distinguish reality from falsehood.
Imagine a world in which we have tools that can generate text in someone’s voice, coupled with other tools that can recreate a person’s voice with just a short snippet of sound, married to technology that automatically creates computer-generated video and images of people, things, and places that appear indistinguishable from real life video.
Soon, by uploading a short photo or video of a person, you’ll be able to tell a tool to make that person say or do just about anything, in just about any setting. In minutes you’ll have video proving that when your older brother was 12 and you were 10, he hocked a loogie off Disneyland’s Dumbo the Flying Elephant, right onto your older sister’s head. It’ll even have that shaky characteristic of video taken while enjoying such a ride.
Another example: imagine a world where you can tell a computer to build a website around a certain topic, and it will do so, populating pages with automatically generated information. In a matter of minutes, you’ll be able to stand up a website that looks as real and fleshed out as a well-established source, with correct or incorrect information as you see fit. You can even give it whatever “published on” date you want.
You can imagine what else might be done with such tools.
So, hold on. Buckle up. It’s about to get weird.
2. Many of these tools should not be called AI. They’re predictive algorithms.
Yeah, I know, it’s nitpicky. But it’s true. The term artificial intelligence implies some understanding of what’s really going on, as well as some range of free will or motivation. But ChatGPT and Dall-E don’t really understand meaning. They only understand 1s and 0s.
To put it simply, Dall-E has been shown millions of images with labels on them. It knows that certain labels (1s and 0s) are often accompanied by certain patterns of 1s and 0s.
The same with ChatGPT. It’s been fed millions or trillions of bits of information with labels, and analyzed what order those words or 1s and 0s are in, and how often those words appear together, etc. But, it doesn’t know what those things mean. It also has no free will or motivation or desire.
It can fool you, no doubt. It can probably pass the Turing Test. You can ask it, “what is a credit union” and it will give you a good definition. You can ask follow-up questions, drilling down and getting more information about the details. It will feel like this computer knows what it’s talking about. It could spout definitions and statistics until you’re blue in the face, to prove that it knows what it’s talking about—but trust me. It doesn’t. It just knows that these 1s and 0s often follow each other in certain patterns.
Ultimately, it’s math, translated into words or pictures. The computer, itself, can only follow the instructions given. It can only do math. It will not do anything unless it’s told to do it.
I suppose, ultimately, this is a discussion about what makes us human, and if that will separate us from future artificial intelligence that appears to act and know and understand. It’s the subject of untold numbers of science fiction novels and movies and TV shows. Some might say that just because the computer’s physiology is different than ours, and it has a different basis for understanding, does not mean that it’s not intelligent. I suppose. Maybe. In the future, when we have general AI.
But for now, don’t confuse ChatGPT or any of its iterations as intelligence. Recognize it for what it is: really great math spanning a limited scope, on really powerful machines.
3. What does all of this have to do with credit unions?
My feeling is that as the world becomes more antiseptic and confusing, person-to-person interaction is going to become more meaningful in business and government settings. When you can no longer tell if you’re on a Zoom call with a person or a digital representation of that person (the digital representation has been trained with that person’s voice and talk and the person has instructed it on how to act during the call), or if what a recorded video portrays really happened—at that point, live, in-person interaction is going be worth a premium. Policy makers will value face-to-face interactions, as will credit union members.
I’m not sure what this means for branch systems that are increasingly moving to video teller machines. Perhaps there will be a backlash at some point, and members will demand real people even for simple transactions.
Maybe there will be a middle ground where the machines become so good that each of your employees will have a digital agent (clearly labeled as such) that performs actions on their behalf, but once a member’s needs become something more than the agent can handle, it switches over to an interaction with the actual employee.
Maybe rote transactions will be handled with a simple user interface—not a representation of a human conducting the transaction for the member. As development continues, what is considered “rote” becomes more and more sophisticated. And, because people are getting used to banking being an interaction with computers, it’ll be fine. But at that point, what will the credit unions do that say their member service is their biggest competitive advantage?
On the Association’s side of things—the advocacy side—when legislators and regulators can no longer tell if they’re talking to a computer-generated bot or not, in-person face time with legislators and regulators will become more important.
Imagine a scenario like this: the state legislature is considering a bill that outlaws lobbying done by robots. Of course, they want this legislation because, regularly, they and their staff members are being contact by a weak AI lobbying against the legislation. This weak AI can even make phone calls and record videos with the faces of real lobbyists. Won’t that be an interesting day? Weird, indeed.
For all scenarios, as digital interaction becomes less expensive and automated, in-person contact will increase in price and in perceived value.
In the end, I don’t know where it will go. I suspect it will get messy and ugly.
And super, super weird.
4. One last thought
When my brother-in-law presented those poems, was it unethical or deceitful of him to say he wrote the poem? Should the person who was moved to tears by the poem have felt deceived or upset to learn that he didn’t actually write the poem?
The gut reaction is yes, until you know that he had to give the tool a bunch of detailed, personal information that made the poetry more personal and relevant—grounded in facts. Those are his original thoughts. He just didn’t have the skill to make a poem out of those thoughts. So, to utilize a tool that made the poem, and then give it to someone—does that make the poem less valuable or genuine or touching?
I’ll let you decide. But for now, here’s a poem about credit unions I wrote for you with the help of ChatGPT. It’s titled “Credit Union Hero” and I gave the following instructions to the tool: “Write a poem about credit unions that talks about how they save members money, and provide better service, and make a big difference in the world.”
I didn’t write it, but the feelings are indeed mine.
Credit Union Hero
In the world of finance, where big banks reign,
There’s a humble hero, with a different aim,
A credit union, it’s name we know,
Where members come first, and profits take the back row.
With lower rates and fees, they save us cash,
And when we need a loan, they don’t make a dash,
Their customer service, it’s always top-notch,
At the credit union, we’re never just a number or notch.
They’re owned by members, not big shareholders,
So when they make profits, it’s not just for their own affairs,
They give back to the community, in big ways and small,
At the credit union, they truly care for all.
With programs for financial education,
They empower us with knowledge and salvation,
So we can make smart choices, and grow our wealth,
At the credit union, we’re more than just a shelf.
They’re not just about money, they stand for more,
With values like integrity, and a heart at the core,
They make a difference, in the world we live,
At the credit union, they always give.
So if you’re looking for a place to bank,
Where you’re not just a dollar sign, but more than that,
Look to the credit union, you won’t regret,
It’s a place where you’ll feel right at home, no threat
No, no, this is the poem I wrote for you.
Wait, no–not that poem. A sonnet. Here, this is the poem, titled: “Utah’s Credit Unions: A Financial Gem”
In Utah’s land of mountains, sun, and snow,
There’s a financial gem that many know,
Credit unions, they stand out from the crowd,
With tech that’s top-notch, their service stands proud.
Their websites shine, with ease of use and speed,
With mobile apps that cater to our needs,
We can bank from anywhere, on the go,
At Utah’s credit unions, it’s all aglow.
Their member service, it’s never a chore,
With friendly staff, that greet us at the door,
They listen and help, with our every request,
At Utah’s credit unions, we’re truly blessed.
And with their rates and fees that are so fair,
In Utah’s credit unions, we’ll always care.
No, wait, I have another poem here . . . .