Artificial Intelligence, Marilyn Monroe and William Shakespeare

Artificial intelligence, Marilyn Monroe and William Shakespeare.

Maybe thirty years ago I worked on one of the first “expert systems”. An expert system is a type of artificial intelligence (AI) system that is designed to replicate the decision-making abilities of a human expert in a particular field or domain. The project's aim was to make managing complex plant, for example a nuclear power station, easier by identifying what alarms were important from a sea of alarms in the event of a significant problem. We built a working prototype but back then computers weren't powerful enough to make the system cost effective.

Now we have computers that are a thousand times faster. There is more computing power on my mobile phone that in a whole 1990s data centre. Consequently, we can do amazing things with Artificial Intelligence.

Let me give you an example – image generation. Within the last few months various different software programs have advanced sufficiently to create images that are very close to life like. You may well have seen, in the newspapers, a deep fake image of the Pope wearing a an expensive white puffer coat. The image is fake (obviously, the Pope wouldn't be seen dead in such an ostentatious outfit). It was created by a piece of software called Midjourney v5.0. The weird thing is that version v5.0 was only released four weeks ago!

Midjourney converts strings of text to images. I have been using it myself to create images that couldn't possibly exist. Included with this article is a picture of Marilyn Monroe. The string of text I used to create the image is as follows - “A picture of Marilyn Monroe, aged to look like she is 70 years old, sitting in a diner.” I hope you agree that the image is pretty realistic. Progress is incredibly fast in this field. Midjourney V6 is expected to be released this summer.

Another example are ChatBots. A chatbot is a computer program that can simulate human conversation through text or voice interactions. Chief among them (this week anyway) is OpenAI's ChatGPT. ChatGPT responds to comments and questions. You can ask it pretty much anything and get a sensible response. Below is part of a conversation I had with ChatGPT.

Me: “Can you write a sonnet in the style of William Shakespeare on the subject of Paella?”

ChatGPT: Sure, here's a Shakespearean-style sonnet on the subject of paella:

Oh, Paella! Thy saffron threads so fine,

Thy rice, so tender, bathed in rich broth's glow,

With meats and veggies, all in harmony entwine,

A feast fit for a king, it doth bestow.

From Valencia's shores, thou art born,

A symbol of Spain's cuisine, oh so grand,

Thou fill'st my senses with aromas worn,

And flavours that spread throughout the land.

Oh, let me feast upon thy golden hue,

And let me savour every single bite,

For in thy depths, there lies a tale so true,

Of culture, history, and lands so bright.

Paella, thy name, it rolls off the tongue,

A dish so loved, forever shall be sung.

I'm sure the Bard could have done much better had he ever tasted Paella and had an hour to spare but the amazing thing is ChatGPT created this sonnet in four seconds!

Why are these computer programs improving so fast? Well they are a type of AI known as learning systems. These are programs that improve merely by being used.

I asked ChatGPT if it was a learning program and this was it's reply: “Yes, I am a learning program. As an AI language model, I am designed to continually improve my responses based on the input and feedback I receive from users. This means that I am constantly learning and updating my knowledge and language abilities.”

Every time a human interacts with AI it improves. And it is improving and becoming cleverer with every interaction. How long will it be before AI is better than a human at doing almost anything?

There are many people, including Elon Musk, who believe that unfettered progress in AI could prove a problem in the very near future. I can't imagine a world in which computer systems become a direct threat to humanity but I can imagine a world where computers perform the vast majority of intellectual tasks. I see no barrier to computers taking over roles in marketing, public relations, design, medical diagnosis, accounting, and even computer programming and law. I believe a computer could be able to ask questions of witnesses, construct arguments and present them to a jury. Hell why stop there? Why have a jury when a computer is free of bias and corruption and is perfectly capable, and probably more reliably, of coming to a just decision on it's own?

But don't take my word for it. I asked ChatGPT the following; “Could artificial intelligence ever perform professional jobs like a lawyer or a computer programmer?”

ChatGPT replied, “Yes, artificial intelligence (AI) is already being used in some professional jobs like law and computer programming, and it is likely that its role in these fields will continue to expand in the future.”

And it's just around the corner!