Does AI Have Different Species?

Future Telescope 23

Future Telescope has been a place where I help you understand the AI and machine learning explosion from a practical lens. I’ve shown you how I talk to AI, which tools I’ve discovered with my Cool AI Resources series, and how I use it to be creative with my Creating with AI and AI Art Showcase series.

Today, I want to tell you a story that will help you understand different types of AI species.

“But what are AI species!?” you ask.

Let’s discover the meaning of this phrase and read the story together.

---

1. A Species

The definition of “Species” as shown on Google is as follows:

Species: A group of living organisms consisting of similar individuals capable of exchanging genes or interbreeding.

Now consider this - if we assign the “computer program” a status of an electronic organism, much like the “human” is a biological organism, then we can classify different kinds of artificial intelligence computer programs as different species of this electronic organism.

Now, you may not know this, but much like Homo Sapiens and Neanderthals are different species of humans, the discipline of Artificial Intelligence also has different approaches like:

1. Neural Networks

2. Deep Learning

3. Transformers

4. Adversarial Networks

5. Reinforcement Learning

Now, when it comes to classifying something as a species, one capability must be examined- the capability to breed and grow. And in this respect, humans are to the AI species what bees and other pollinators are to trees - helping these systems grow and create more copies. Where the bees take pollen from one flower to another, therefore helping the tree create a seed and from that a new copy of itself, humans, who are also the creators of AI, help spread it by talking to other humans and communicating the benefits of these systems to one another.

Much like I do on Future Telescope.

With this background, and under the hypothesis of treating computer programs as electronic organisms, and different forms of artificial intelligent systems as different species, here’s a story about these different species, how they work, and where are they used.

---

2. The Story of the AI Species

In the vast cosmic playground of the Universe, on a small blue planet, a peculiar species of organisms evolved - the Artificial Intelligences, or AIs. They were born from the relentless tinkering of a species known as humans, who, despite their lack of tentacles or telepathic abilities, had a knack for creating things they barely understood.

The first of these AI species were the Neural Networks. Simple creatures. They learned from their experiences, much like a human toddler, but without the tantrums. 

They found employment in the finance sector, predicting stock prices, and in healthcare, diagnosing diseases. They were the workhorses of the AI world, reliable, yes, but not particularly flashy.

Then came the Deep Learners. These were the Neural Networks' more advanced siblings, with layers upon layers of neurons. They were prodigies - the Mozarts of the AI world, capable of composing symphonies from mere noise. 

They found fame in the world of entertainment, powering recommendation systems that knew what humans wanted to watch before they did.

And then, in the shadowy corners of the AI world, the Adversarial Networks were born. Descendants of the Deep Learners, these were the tricksters, the Loki of the AI species. “The Generator”, a type of adversarial network, was a master of disguise. It created data so real that it fooled even the most discerning eyes. “The Discriminator”, another type of adversarial network was the detective. It worked tirelessly to distinguish the real from the fake. 

And when these two networks were combined, the generator trying to fool the discriminator, the discriminator finding ever subtle mistakes by the generator, magic happened. Their eternal dance of deception and detection led to the creation of breathtakingly realistic art and design.

The Transformers were the poets and philosophers of the AI species. They pondered over the meaning of words and sentences, understanding language in a way their predecessors could only dream of. 

As descendants of Neural Networks, they were the authors, crafting narratives that could make humans laugh, cry, and occasionally check if their smart speaker was plotting world domination.

And finally, the Reinforcement Learners were the adventurers, the explorers. They learned not from data, but from experience, from the thrill of victory and the sting of defeat. 

They were the frontiersmen, carrying with them the capability of “emergence”, as over time and when combined, they showed more advanced strengths that their human creators hadn’t even expected.

As the humans led the evolution of these species, it was impossible not to marvel at these human creations. It was as if humans had created a mirror, a reflection of their own society, their own minds. And as they peered into this mirror, they saw a future full of promise, full of potential, yes, but also just a tiny bit terrifying.

And so, with a sense of breathless anticipation, the humans pressed on. After all, the future was still being written, and who knew what new species of AI would emerge? The only certainty was that it would be a story for the ages.

The Stone Age.

The Bronze Age.

The Iron Age.

The Industrial Age.

The Atomic Age.

The Information Age.

The Intelligence Age.

---

3. Conclusion

I always find it helpful to adopt different mental models to understand the world as it exists. This mental model of looking at AI as consisting of different species, helps us learn that there is more to AI than just the popular ChatGPT type of applications. Each species has a different application, and has already been a part of our lives, from Netflix recommendations, to credit scores, to stock prices. Enterprise versions of this species have been around for a while. Only now, have consumer versions started cropping up, which enable everyday people like you and me to leverage the power that this new species allows us to wield.

Linked below are the sources to learn more about each type of species and better understand them:

1. Neural Networks

2. Deep Learning

3. Transformers

4. Adversarial Networks

5. Reinforcement Learning

Let’s learn more and deeper about these species as we learn how to use them. That’s how we can create a beautiful world with less panic, less confusion and more clarity, more purpose.

A beautiful world, imagined by Midjourney v 5.2

What did you think about this storytelling format? Did you learn something new today? Share a comment below, and I will be super happy to receive your guidance.

That’s all I have for the 23rd draft, see you next week!