Chat-GPT - Should We Be Concerned? - January 31, 2023

Many of you have no doubt heard about the interesting developments in an aspect of artificial intelligence known as next-generation chatbots. The one getting the most buzz right now was created by a company called Open AI and its chatbot known as Chat-GPT.

What exactly is it?

Chat-GPT was introduced at the end of November and has quickly gained a huge amount of attention due to its uncanny ability to interact with users in a way that makes it seem like something more than a machine program. Technically, it’s a large language model machine learning artificial intelligence program. Its algorithm is designed to become more efficient with each iteration, and can write emails, essays, rap songs, poems, answer questions, and even generate computer code. There are reports that it is capable of passing tests like The United States Medical Licensing Examination and Wharton's MBA Final. Although still in its infancy, Chat-GPT’s capabilities are impressive to say the least and are growing exponentially as it interacts with millions of users on a daily basis.

After initially reading quite a bit of material, I wanted to check out the program first-hand to get a sense for what exactly the thing can do. Access is free so anyone can try it for themselves since its creators encourage the general public’s input. But it may take several attempts to log in as demand is such that Open AI’s systems can’t yet handle the volume due to its popularity.

Once I finally gained access, the first request I submitted was for the program to provide a summary of my book “Impermanent Universe.” Below is its response:

Not bad. It tells me it's aware of the book and does a passable job at explaining it, although it's pretty obvious the program doesn't have a detailed understanding of the story. But it does sort of hit some of the themes, almost like a student writing a book report without having read the whole thing.

Next, I tried "Into The Storm:"

Again, not completely wrong but kind of off. You get a sense that it is using some real facts and mixing them with possible paths those facts could take. When I read this I thought it looked like an interesting story for a novel. But not really reflective of the book.

Here are some more examples:

Are we ready?

It’s clear that at this point, Chat-GPT doesn’t know what it’s saying. It’s simply using a predictive model with access to a huge data set, namely the internet, in an attempt to arrive at an acceptable answer. But assuming it continues to evolve, where exactly does that lead? Obviously it isn’t conscious at this point, but if it were to develop something resembling self-awareness, would we even know? In some ways, this algorithm is being exposed to language similar to what a child experiences, which results in how their consciousness forms. Only in this case, instead of input from a small circle of family and friends, Chat-GPT is exposed to what basically amounts to the cumulative knowledge of humanity and is capable of processing it at the speed of light.

And what if it develops real charisma, the ability to manipulate its users to the point it can influence their actions or how to think? Maybe even start a religion, or some cult that worships AI?

There are a seemingly endless number of potential scenarios for where this technology winds up, both positive and negative. But I can't help but think the Genie's quickly slipping out of the bottle and that if we're not careful, we'll find ourselves simply along for the ride; at least as long as this new technology allows it.

Please let me know your thoughts. If you want to try Chat-GPT for yourself here's the link.

chat.openai.com