The AI (R)Evolution

In nature, frogs are fascinating creatures. They incubate their eggs, spawn their young, and then transform from tadpoles into frogs within the space of a few weeks. Thousands of them push off and take that big leap (of faith?) into the unknown in different directions. 

As you watch, a deep and irrevocable transformation occurs, shifting shape, form and function in the blink of an eye. Rapid. Exponential. Transformational. 

Artificial though it may be, AI’s journey resembles the frog as it unfolds, a silent ooze seeping into our lives, leap-frogging from function to function. But perhaps the most marked similarity lies in its evolution and what that means for us. 

We started off on the straight and narrow. 

Yes, I’m talking about AI that was designed to conduct a specific task or process, within a specific context and vertical. We call this “narrow” AI because it is tailored to play a specific role e.g. facial recognition, and it sticks to that. 

But remember, digital darlings, AI is ultimately created by us - the fallible human being, and straight and narrow is so boring after a while. Enter a touch of the deviant, and we have the next baby (oops!) which is Artificial General Intelligence.

This is the generalist, the one which can connect A to B to C, based on its ability to create patterns within a large dataset. Practice makes perfect. The more frequently AGI performs a specific action, the better it gets. Just like your brain. No wonder it is called a neural network, no?

Then there’s Generative AI. This is the showstopper, the face we all recognize. ChatGPT made the headlines. Gemini has two faces. But the bottom line is that Generative AI is a dynamic beast. Think of a smart cockatoo, which can mimic what you say and do. The proverbial cow that chews thoughtfully on its feed and then regurgitates that cud in a giftwrapped box.

But now we have a new agent. A smooth operator if there ever was one. This one is called Agentic or Super AI. Agentic is trained on SMART goals - and it’s a regular smarty pants.  Agentic is the octopus of AI. Shrewd, smart, and able to make decisions on the fly, navigating different scenarios and situations even as they crop up. It’s autonomous, and perhaps to some degree, amorphous, because it isn’t “narrowed down” nor is it reactive. It’s responsive - and its proactive.

IBM describes it well:

“Agentic AI describes AI systems that are designed to autonomously make decisions and act, with the ability to pursue complex goals with limited supervision. It brings together the flexible characteristics of large language models (LLMs) with the accuracy of traditional programming. This type of AI acts autonomously to achieve a goal by using technologies like natural language processing (NLPs), machine learning, reinforcement learning and knowledge representation. It’s a proactive AI-powered approach, whereas gen AI is reactive to the users input. Agentic AI can adapt to different or changing situations and has “agency” to make decisions based on context.”

(Agentic AI vs. Generative AI | IBM)

But like the Octopus, is it also a thief, stealing your gem of an idea? 

The first point of departure for Agentic AI begins with “perception” – one that is based on data that it draws from its ecosystem. 

The cookie has crumbled, my digital darlings. The Octopus a.k.a. Agentic AI, sucks up your deets every time you tap your screen. It scores your number from databases, scrolls sensors and tracks APIs to know all there is to know about you.

Then it crunches it up and culls out the relevant information that it feels it needs. It uses NLP (Natural Language Processing) as well as computerized vision to do its own little psychoanalysis. 

The one thing I cannot deny about Agentic AI is that it doesn’t dither. I also like that it thinks things through and responds. If you nudge it along in the right direction, it gives you well thought out responses that make the opposite person feel heard. And it tells you how it arrived at that conclusion.

Based on the inputs it processes, it sets its goals, develops a strategy to achieve them, and then takes decisive action at a breathtaking pace. And like any intelligent (artificial or otherwise) being it does a thorough evaluation and introspection exercise to learn from itself.

Press pause. Did you take that in? Because if Agentic AI shapes up to be what it wants to be, you may have an AI clone, sooner rather than later.

How many times have you thought to yourself, if only I had another person like me to do everything. Well… NOW YOU DO.

And that’s where blockchain can set the boundary, between the “real you” and the “AI avatar”. Allow me to explain this. 

Blockchain lets you be as private or as public as you want to be. Think of it like the volume on your music. You can crank it up to complete transparency or dim it down to total opacity. If there are aspects you’d rather withhold from AI’s grasp, blockchain may be able to help you do just that.

That being said, AI is pretty pervasive. So, what blockchain can also help you do is establish your ownership and IP. Blockchain creates an impregnable record to establish the rights of a work and attribute them directly back to their creator, before AI can purloin them.

The bottom line is that Blockchain builds trust in the system, precisely where AI often tends to fall short. AI’s hallucinations could be mitigated by blockchain verifications and authentications. 

In the media & entertainment world, blockchain could also flag off where AI is cloning talent or creativity without the artist’s permission and pre-empt piracy.

In an artificial ecosystem, blockchain keeps the intelligence real.