Artificial Intelligence and Why You Should Care
Technology is evolving so fast that by the time you finish this sentence, someone has probably invented a new AI that writes better intros than this. While Artificial Intelligence isn’t exactly a new concept (it’s been lurking around since the 1950s), it wasn’t until OpenAI dropped ChatGPT-3.5 in November 2022 that the world collectively lost its mind. Suddenly, AI wasn’t just some sci-fi fever dream — it was writing our emails, passing medical exams, and gaslighting us about whether it really is sentient. So, what exactly is AI, and why is everyone (from tech bros to your grandma) talking about it? Let’s dive in.
What in the Algorithmic Hell is AI?
In the simplest terms, Artificial Intelligence is when computers act like they have a brain. “Wait — so my dishwasher is intelligent now?!” Well… yes? But also, absolutely not. The problem is that "intelligence" is a pretty blurry concept. If you really wanted to, you could argue that your dishwasher is a form of AI — after all, it makes decisions, follows programmed instructions, and occasionally beeps at you in frustration. But unless your dishwasher starts writing poetry or plotting world domination, it’s probably not what we mean when we talk about AI today.
AI Comes in Many Flavors: From Slightly Smart to World-Dominating (Someday?)
AI isn’t a one-size-fits-all deal. Just like humans, some AIs are geniuses, and some can barely function without supervision (looking at you, autocorrect). You wouldn’t ask a toddler to solve climate change (unless you enjoy finger-painted disaster plans), so why expect Siri to answer life’s biggest questions? AI has different levels of intelligence, ranging from kinda useful to slightly terrifying:
- Artificial Narrow Intelligence
- Artificial General Intelligence
- Artificial Super Intelligence
aka our inevitable overlord
Let's give them a closer look.
ANI
These are systems with limited capabilities. This type of AI is great at specific tasks, and terrible at everything else. (Think: spam filters, Netflix recommendations, and that AI that keeps suggesting you buy stuff you already own.)
AGI
An AI that learns, thinks, and adapts like a human — with better memory and fewer existential crises. We’re not there yet, but at the rate AI is advancing, who knows? By the time you finish this article, we might just have one. (No pressure, scientists.)
ASI
This is the level where AI outsmarts all of humanity combined. It could cure diseases, reverse climate change, solve world hunger along with all the other issues that politicians like to spend so much time talking while doing nowhere nearly enough about.
"So that's it, right? Just the three types?" Not quite! AI also has sub-types, which are useful to know (especially if you want to sound really smart at parties), so let's dive into them.
- ANI:
- Reactive Machines: These systems have no memory and only respond to very specific tasks, kind of like an NPC stuck in a dialogue loop.
- Limited Memory: I think the name speaks for itself but in case you missed it, this type of AI has — wait for it — limited memory. That means it can learn from past data, but only within a fixed scope. Think of it like a goldfish with a really good short-term memory.
- AGI:
- Theory of Mind: AI that actually understands emotions, intentions, and beliefs. In other words, it’s the moment AI goes from smart assistant to therapist who knows why you’re sad before you do.
- ASI:
- Self-Awareness: Another type with a self-explaining name, but just so we are on the same page: this AI has full awareness of itself, its surroundings, and probably the fact that we’re all nervous about it.
So, where do we stand with AI right now? We are still in the ANI category, but maybe not for long!
Now that we have covered the types of AI that are currently known of, how about we move on to the mechanisms behind it, shall we?
How Do Machines Think? (Or: Why Your Toaster is Smarter Than You’d Like)
So, why does AI seem so smart? What’s happening behind the scenes when it “thinks” and “reasons”? While explaining all the technical intricacies of Artificial Intelligence’s… intelligence (yeah, yeah, forgive the repetition — blame the English language) would take way more than one article, I can at least break down some key concepts that’ll help.
Pattern Recognition
One of the core concepts behind AI’s thinking process is pattern recognition. When you train an AI model, you feed it a ridiculous amount of data, and it starts noticing patterns.
For example, imagine you train an AI using 1,000 pictures of cats. What does the AI do? It studies characteristics like size, color, breed, body parts, and behaviors until it forms an idea of what a “cat” is. (For instance, if you train an AI on funny YouTube cat videos, it’ll probably conclude that cats love knocking things off tables and that orange cats operate on pure chaos.)
Have you ever noticed that ChatGPT gives you better responses if you add "please", "thank you", and "I will tip you if you" to your prompts? No? Well maybe that's because you should try being nicer to it.
Turns out, multiple research studies suggest that LLMs perform better when users are polite. But don’t panic — this isn’t because AI is developing emotional intelligence (yet). The real reason? The models were trained on human data, and humans tend to perform better when treated with kindness.
Would you not agree that there is a significant difference between these:
- "Wash the dishes"
- "I'll give you $5 if you wash the dishes"?
AI recognizes that politeness is often associated with better performance and, well… acts accordingly.
Learning from Experience
Another key concept in AI is its ability to learn from past experiences. If you play thousands of chess matches against an AI, it won’t just memorize moves — it will adapt, recognize strategies, and get better at playing.
"So THAT’S why I never win against the computer!"
Exactly. And now that you know, I hope it makes you feel just a tiny bit better. :)
Probabilities and Predictions
Unlike what you might think, AI doesn't always know the answer, it usually makes an educated guess based on probabilities. Kind of like when you run into someone you definitely know, but their name is gone from your brain. You take a wild guess — "Hey… BRIAN?" — and just pray they don’t correct you. So, when your phone suggests the next word while typing, it's predicting what is the most probable thing that you will say next.
Well, we’ve covered quite a bit, haven’t we? I hope that by the time you are reading this you have some basic understanding of key concepts of AI and are seriously considering adding things like "please" and "thank you" when talking to ChatGPT.
It was my pleasure writing for you and I hope that you have enjoyed learning just a little bit more about AI!
Written by
Anastasiia Ovchinnikova, who has always been nice to ChatGPT
Top comments (0)