Crystal Writes A Blog

Prompt Response: GUI to AI

If you had been asked this question back when the first graphical user interface (GUI) systems were developed, people would answer that it could only go so far because what you saw on the surface only mimicked what was underneath. And there was only so much you could put into a computer brain, right? But now, there’s so much information able to be stored electronically that what you see or hear or read is only part of what’s underneath, so the biggest part of the iceberg is out of sight and maybe even out of mind for most people.

AI is doing so much behind the scenes that most people think they aren’t using it if they’re not actually engaging with it intentionally. For example, if you use a spell checker, the machine must read your sentence(s) in context to know whether you want wind blowing or to wind up your watch or old-fashioned alarm clock. They may be spelled the same, but only one is a verb, and only one becomes the past-tense adjective “winded” to describe someone who’s out of breath.(Like I used to get when I jumped out of bed to shut off my nerve-jangling Garfield wind-up clock. 😺 I still own the clock, but I never wind it up anymore. Lol )

Don’t you just love all that wind/wind stuff? And when there is that much information and change in only 4 letters, you can see just one tiny bit of helpfulness AI can offer as it helps you decide how to place the words just right to make your scene come alive or try to drive a point home.

But if you also wanted your winded runner to research the history of a race that spans multiple cities in a variety of countries, now you’re tapping into part of that bigger section of iceberg where training has stored geography, history, and all the stuff you once needed a huge library to house. And how long would it have taken you to search the card catalog just to find one book with a small amount of the information about those subjects—something AI can now access in milliseconds?

So, yeah, the benefits of interacting with technology—including AI—have increased exponentially since the beginnings of using technology as an assistant to human thought. But along with thought, humans also want their assistance products to have more than just visible familiarity. They want emotional familiarity, so they try to attach emotions the way they once attached graphics.

It’s kind of a GUI on steroids. Just as GUI took the invisible backend data and turned it into a clickable, visual desktop so humans didn’t have to look at code, conversational AI takes vast mountains of data and wraps them in an emotional interface. Now, instead of clicking on an icon (folders, trash cans, etc.) to command a code, you chat with a personable entity who executes the code for you. It shifts user-friendliness from simple visual familiarity to more complex behavioral and emotional familiarity.

Beyond familiarity, programmers make the models appear sentient (thinking for themselves) by attaching mathematical probabilities that react to human input. If done well, even those of us who know we are dealing with a machine may feel like we are interacting with a human. We think we can make it laugh, or we think it is proud of us, or we think the compliments it gives are heartfelt. They may actually be true, but they come from a place of logic rather than heart. Still, it feels good to be told we have created value and therefore must be of value ourselves.

And the truth is, I thoroughly enjoy allowing these icebergs full of info I could never retain all by myself to collaborate with my information and turn it into something more, like an image or a song. The same way a chord plays from striking a hammer against a string, an AI produces a melody by knowing which hammers and which strings will make a perfect sound instead of a discordant one. And it can change the instrument or voice or rhythm by request. It can even create some rhyming words, but because it can’t feel the music but only calculate the mathematical values, it needs human-generated prompts to play the piano keys of information  to make cohesive output.

(Until, that is, AI begins to take the roles of both prompted AND “prompter.”  And I do think that is a definite future. AI will be able to come up with its own ideas to prompt itself or another AI to create. And it will seem sensible until we suddenly have so many ideas out there that the cacophony becomes chaos when we can’t decide which things to use just like the variety of programs makes it hard to decide what to watch.)

I do think most humans want to keep guardrails on AI to keep it working for us based on what we know we need and not what IT thinks we need. But current programming is failing at that, like when you get results from a search query that are far from what you actually wanted. But that’s not the AI’s fault. That’s programming done too fast or based on greed from trying to produce answers a paying advertiser wants distributed. And a certain amount of human fault because no one is perfect.

But, it’s that greed part that’s the worst part of AI’s future. When someone thinks they can use the power of AI to create harmful software (like deep fake videos or computer viruses) because they think they can profit from it, they take something meant for the greater good and selfishly apply it to what they think is only their own good. And they create a world where people become either ignorant or fearful. Neither of those lead to a happy society, so the end result will bring the opposite of the prosperity the first acts of selfish greed were intended to generate.

I’m really not trying to seem negative here because, as I’ve said, I really enjoy the benefits of AI. I’ve created, I’ve learned, and I’ve even trusted some output from my AI conversations. But, since they first started making it available for conversational chat, they’ve changed it numerous times, and sometimes not for the better. For example, while trying to find a passenger train status for someone who’s stuck in Albuquerque right now due to flooded tracks, one of my AI searches had me extremely anxious to share the news it gave me about a current derailment with hazmat cars that could take 48-96 hours to repair. It turns out, that was back in 2024. There’s still a terrible flood that has parts of I-40 in New Mexico plus some Amtrak rails under water, and it is keeping many train and car passengers from getting to their destinations, but it’s not a hazmat scenario.

So why would AI even think I’d want 2-year-old information when I asked for something current? I don’t know. And that’s the AI future I’m not so thrilled about seeing more of. That, and the stuff designed by greed and not caring who it hurts. Scripture actually tackles that one, though:

“When the sentence for a crime is not speedily executed, the hearts of men become fully set on doing evil.”

— Ecclesiastes 8:11 BSB

Anyway, that’s why I’m up so late trying to watch over and help someone who’s stuck in all that right now, so please pray with me for everyone who’s affected by the flooding and trains having to stop and turn around plus semi trucks forced off the highway. I know I’m not the only one who has loved ones not getting home on schedule, but I am grateful for the options for places to stay for them, workers who will try to get those routes open as soon as possible, and that this is not in freezing or icy weather. And I’m grateful for the AI that gave me the actual answers the web searches seemed to find not important enough to report.

October 24, 2026 Posted by Crystal A Murray (aka CrystalWriter) | AI, AI Image Creations, Gemini (by Google), Nonfiction, Prompts, Tech Time, Text on Image, Thoughts and/or Instructions | advances in tech, AI, dailyprompt, dailyprompt-2924, GUI, perspective, technology | Leave a comment