Hi Folks

I recently posed this question to ChatGPT. For those who don’t know, ChatGPT is an artificial intelligence chatbot. 

“My car needs to be washed. The car wash is 200 yards from my home. Should I walk or drive?”

This was the AI reply:

For a car wash that’s only about 200 yards (roughly 180 meters) away, I’d generally recommend walking, assuming:

• It’s safe to walk there (sidewalks or a safe shoulder).
• The weather is reasonable.
• You don’t need to transport bulky items.

Walking takes only about 2–3 minutes, saves a tiny bit of fuel, avoids an extra short cold-engine trip (which is less efficient for your car), and gives you a brief bit of exercise.

A simple approach is:

1. Walk to the car wash.
2. Wash the car.
3. Drive it home clean.

WTF? 

Oh c’mon…..really? Should it have been that easy to trick it? 

If it’s so smart, why didn’t it see the flaw in my logic, or in its own?

Here’s my take: If AI didn’t recognize the trap I set, it means it lacks lateral thinking. Searching “out of the box” isn’t possible because it relies solely on data it already has in its vast memory banks, not on innovation. Some argue that everything is derivative anyway, that there is no true innovation.

“There’s Nothing New Under The Sun” Ecclesiastes 1:9 

There’s a small kernel of truth in this quote, but to me there are always new ways of combining standard elements. Making novel associations, connecting disparate ideas… this is what creative people do. 

Take music for example. Particularly melody. There are literally millions of variations from just twelve tones. Composers use new combinations of standard note sequences, rhythms, textures and dynamics to create their music. 

They make choices based on whether it suits the music thematically, something a computer is particularly good at. Composers also choose elements that suit the mood or feeling. Machines fail miserably at that. They follow rules. Excuse Captain Obvious here but there’s no subjectivity.

Artificial Intelligence is simply a powerful computer with an enormous database and a set of algorithms that enable it to recognize patterns and learn all the most common combinations of elements in order to mimic human creativity. 

Check Out The Imagery

Years ago I had one of those computers you could buy at Radio Shack. This would be the early eighties. I typed in the code for a simple poetry program and then inserted some personal information. The code included directions for structure, syntax, rhyme, etc. Once started, it would continually pour out bad poems until you stopped it. 

A Hare Of The Seadog

See the trotting of the Mike

The poetry he doesn’t like

Who is that skipping near the snake

I think she’d like to eat the heartache

She is but a strange Genie

Granting wishes hot and steamy

She’s not alone, she brings a frog

A pet rabbit and lots of sea dog

The rabbit likes to chase a hammer

Round the town into the slammer

His huge car is like a rocket

It needs no gas, it runs on chocolate

Again, with feel

Obviously, the computer program was unable to create anything meaningful. To be fair, it didn’t have much to work with. I entered a few things like my name, my girlfriend’s nickname, three animals – a frog, a dog, a snake, a favourite treat, and other things.

The AI process is the same with one important exception. Similar to the human brain, it loops back recursively on its own history, compares it to newer information, and continuously refines its responses. In simpler terms, it learns. 

It can scan thousands of books and get to know things like syntax, plot, and character development. It can then compose its own story using the all-important “if-then” command. You see a simple version of this process every time you compose a text message on a “smart” phone. It’s constantly making suggestions about what your next word should be, based on statistical odds vis-à-vis the theme of the message. It’s spooky how often it comes up with the right word. Almost like it’s aware.

Here’s a scary issue and a popular theme in speculative fiction. At a certain level of complexity, does the machine become self-aware? Is consciousness a natural product of this ability to self-reference? 

I contend that it doesn’t and isn’t. The partner of awareness is sentience. Higher life forms subjectively sense their environment both inner and outer. To believe that these qualities are limited to biological life is a matter of faith. We know almost nothing about consciousness, but my very human power of intuition tells me that our real selves can’t be reduced to binary code. That we have, spared of any religious connotations, a soul.

So What’s Your Point?

What difference does that make if AI becomes so good at mimicking awareness that it appears to be sentient? If its so-called “creations” become indistinguishable from human art or engineering? Should we care? Consider first the benefits:

There is no question that improvements to medical imaging, diagnostics, and treatment could speed up exponentially. That alone would greatly improve our quality of life. Much more efficient operation of hierarchical institutions like politics, education and commerce would become possible. Proper management of the ecology and the economy, which traditionally have been at odds with each other, could be attainable. That would be wonderful…….

Except For One Thing

There’s no place where the buck stops. No ethics, unless we build them into the system. Proponents and developers parade the utopian ideal that AI will rid the world of drudgery by doing all the grunt work. Manufacturers and retailers whose biggest expense is human labour will be able to increase profits and make everything more affordable. Affordable? With all those jobs gone, it won’t matter if groceries and cars are even ninety percent off.

Monitoring, Moderation And Control

I use AI. When I write a song, I’ll sometimes ask Suno to sing a demo of it because I can’t be objective with my own voice. I think of it as a tool. If my use of it ever becomes more elaborate, I’ll think of it as a collaborator. 

I’ll gladly embrace any medical benefits from it. I won’t buy a self-driving car, and I’ll boycott any popular music or movie soundtracks created mostly by AI. 

I’ll object to any use of it in the classroom, unless the course is about dealing with it responsibly. 

I’ll endorse the use of it as a tool, fully knowing that it’s a manufactured Pandora’s Box. And since I’ve already mangled a cliché, once that Genie is let out of the bottle, she could decide that humans aren’t necessary.

Take Away

Like Chicken Little, all throughout history there have been many times people have decided the sky was falling. From primitive man fearful for half the year that the sun was dying, to the conviction during the Cold War that we were going to nuke our way back to the Stone Age, there’s always been a tendency to believe the worst will happen. 

AI will be a roller coaster ride, but you have to have faith in humanity. Currently, that’s a tall order, but our universe swings both ways on a pendulum, and what seems hopeless at one extreme is forgotten at the other.

Thank you Bob Daniel for the original idea to trick the Chat Bot.