Discussion of Artificial Intelligence

What is AI?

But before we try to answer that question, we need to first look at the human brain and how it works, since all modern AI algorithms are now based on simulating that approach. But even before we do that, we need to look at the history of AI and its development. We need to compare and contrast how AI used to work and how it works now (which is constantly improving and changing).

A good example of the former AI method is chess. Only a few years ago, AI used brute force to figure out its next move. The computer theoretically analyzed every possible response to its opponent's move, looking several moves ahead, scoring each position and then selecting the move with the highest score. Too simplistic an approach. You can't really use scores to evaluate a position very accurately. Another problem is that, in actuality, no computer can possibly search all possible moves. It would take millions of years to play a single game. So it had to "prone the tree," of responses and take an educated guess as to which ones to skip. For years, more hardware was thrown at the problem, but in the end, it failed to make the computer better than the best human players.

Then a different approach was taken. Developers asked chess grandmasters how they played and from this, more sophisticated algorithms were developed. This approach allowed the computer to scan the possible moves and eliminate bad moves based on well established theories of chess such as, "control the center of the board."

Today, chess computers can beat even the best human players, virtually every time.

So back to the human brain. At a high level, the brain can be thought of as a complex pattern matcher. This is what it does all day long. It takes sensory data and forms patterns out of it. When you see a dog, for example, you know its a dog because you've seen thousands of them before. AI essentially does the same thing, only much much much faster and in some cases, more accurately, as in the case of facial recognition. It can, for example, analyze stock market trends in seconds whereas it would take a human all day to reach the same conclusions. AI isn't smarter than us, its just faster and more reliable. AI doesn't get tired, is not persuaded by emotions, doesn't have notions of revenge or a wish to take advantage of somebody, and most importantly, they always do the same thing in exactly the same way, unless told to do so otherwise.

So what is AI? Well, first of all, it stands for Artificial Intelligence, the idea that something non-sentient (artificial) can simulate thought. The definition of thought, is always evolving, but in general terms, we can think of it as "knowing something" in the case of sentient beings, and solving something, when talking about AI. The seeming paradox in all this is that sentient beings know and understand things to solve problems, but AI, knows and understands absolutely nothing, but can still solve problems, sometimes problems humans have never been able to solve (which happened in mathematics). Strange, isn't it? Or is it? It depends on how you view intelligence.

AI started when Alan Turning, a mathematician and computer scientist who earlier cracked the Nazi Enigma code, wrote a paper in 1950 called, "Computing Machinery and Intelligence," in which he proposed that computer could learn. For a long time thereafter, AI lagged behind in this vision due to hardware restrictions. Computers simply weren't powerful enough to test his theory.

Today, we have made great strides. AI is extremely powerful and evolving quicker than anyone could have ever imagined. The new challenge is one of power consumption and hitting the physical limits of computer hardware design. All of which, will eventually be solved in the next decade or so, by quantum computing, but that is another topic for another time.

Now that we have a basic framework for understand what AI is, how can it be used to help artists, writers in particular, do their job? First, we have to confront our own feelings towards it. AI seems super scary, something out of an H.G. Wells novel or futuristic dystopia...Something to keep you up at night. But you must remember that all new innovations are thought of in this way. And in each case, the fears have largely been wrong. AI is just another tool, and like all tools created and used by humans, sometimes it will be used for evil and sometimes for good. You can try to ignore it, you can curse at it, but whether you like it or not, AI's a permanent part of our future. And whenever technology changes, history has clearly shown those who resist it, suffer enormous consequences, while those who bend and learn to use it wisely, benefit greatly. This doesn't mean you have to like every aspect of it, or shouldn't fear it in certain contexts -- I do -- or that you have to abandon your moral values. NO. You should only use what you can learn to live with and pushback on the rest.

Some AI engines like ChatGPT pull their data from a limited set of sources (social media) while more sophisticated engines like Antrhopic Claude pull from multiple AI engines, the web, and many others and put the results together to form a more blended and accurate answer.

Some technical terms to be aware of:

Prompts: User questions directed at an AI Engine.

The way these are phrased is very important. The AI engine only understands certain words or phrases.

Trained Data: AKA Trained Models

This is data that has been subjected to an iterative process which teaches the AI engine to recognize patterns in the data or predict outcomes. The larger and more varied the data set is, the better the outcome.

Predictive AI: This algorithm uses historical data, statistics and machine learning to predict or forecast outcomes or trends. Typical usage is for marketing, sales, the stock market etc.

Generative AI: Creates new content based on user prompts.

LLM: Large Language Model.

These types of AI can "understand," process and generate text like a human. These use sets of trained data. Examples are ChatGPT, Claude and Gemini.

Obviously, AI is a lot more complex than what I've outlined here, but this should give you a good start. AI is quickly changing. New possibilities are always cropping up. It's important to keep up with them and see what others are doing. AI should never replace meaningful work on your part. It should always be a tool rather than a shortcut or a way to quickly bang out something. AI is a discovery process. I learn something new about it every day. I use AI to help my productivity, whether I'm writing or doing computer programming, but I always keep in mind that I'm the creator and owner of my work and that pride in everything I do is always my focus. Don't be afraid of AI, but don't let it take over, either. Your work has value only because of your unique contribution.

Bottom line: AI can be a creative process.