Is the Artificial Intelligence Future Already Here?

Should we be thinking more about artificial intelligence as a society? Will ethical issues arise too abruptly for society to have sufficient time to resolve them?

Maybe.

At the very least, advancements in AI will change society to some degree, especially the implications for the economy.

On the one hand, AI can improve productivity and efficiency, allowing businesses to automate many tasks and processes that were previously performed by humans. This can lead to lower costs, higher profits, and increased competitiveness for businesses that adopt AI technologies.

Additionally, AI can enable businesses to develop new products and services that were not previously possible, opening up new market opportunities and creating value for customers.

For example, AI can be used to analyse vast amounts of data and identify patterns and trends that can help businesses make more informed decisions and gain a competitive edge.

But AI can also have negative implications for the economy. AI technologies can replace human workers, leading to job losses and potentially higher unemployment. This can create social and economic challenges, such as income inequality and the need for workers to retrain for new roles.

Those last four paragraphs, by the way, weren’t written by me but by an AI chatbot called ChatGPT launched last week by OpenAI, a San Francisco AI research lab.

ChatGPT only launched last Wednesday but already has more than one million users. Apparently, that’s a faster uptake than Netflix, Twitter, Facebook, and Instagram.

Many have made breathless declarations like ‘ChatGPT might be the most incredible tech to emerge in the last decade’ and ‘one of the most disruptive technologies ever created.’

Is that true?

Chatting to AI — ChatGPT

ChatGPT is a prototype AI chatbot capable of generating ‘human-like’ text.

I asked this chatbot to explain how it worked:

‘ChatGPT is a large language model that has been trained on a massive amount of text data. It uses this training to generate human-like text based on the input it receives. When given a prompt, ChatGPT uses its internal algorithms to generate text that is appropriate to the context of the prompt. It does this by analyzing the input and making predictions about what words or phrases are likely to come next, based on the patterns it has learned from the training data. In this way, ChatGPT can generate text that is coherent, relevant, and appropriate to the given context.’

When I prodded further, curious about the prediction of ‘what words or phrases are likely to come next’, ChatGPT replied:

‘ChatGPT uses a combination of machine learning algorithms and natural language processing techniques to make predictions about what words or phrases are likely to come next. It first analyzes the input text to identify patterns and structure in the language. It then uses this analysis to make predictions about what words or phrases are likely to come next based on the context of the input.’

Raised on a trove of text data, ChatGPT see patterns and extrapolates from them. The public launch will also add data in the form of user feedback.

Responses deemed irrelevant or incoherent by readers will tweak how ChatGPT ‘predicts’ future words and phrases.

Big, diverse data is the grist to its mill, and it can pump out some interesting stuff.

Here is ChatGPT as a maths companion:

It gives you the answer and explains the logic, something that isn’t immediate with a Google search:

In a similar STEM vein, programmers took to Twitter to express how well ChatGPT handled coding queries.

The chatbot can apparently debug lines of code, too:

Using its strength in calculations, some users recruited ChatGPT as a dietician — prompting it to create a weight loss plan with calorie targets, meal plans, and even a grocery list.

Unfortunately (fortunately?), ChatGPT can’t yet sustain deeper human relationships…

Here to chat and create, but no expert

Neither is ChatGPT construed as an expert.

It can generate human-like text from a constellation of data but cannot (yet) bring its data wrangling beyond the derivative.

When I prompted it to generate an investment idea on tech stocks, ChatGPT responded with this:

‘One potential investment idea for technology stocks is to focus on companies that are developing cutting-edge technologies or innovative products and services. These companies are likely to be at the forefront of the industry and could benefit from strong demand for their products and services, as well as potential advantages over their competitors.

‘For example, an investor could look for technology companies that are developing new technologies in areas such as artificial intelligence, virtual reality, or the Internet of Things. These technologies have the potential to disrupt existing industries and create new market opportunities, which could drive strong growth and profitability for the companies that are developing them.’

Not groundbreaking but covers the basics.

Going further is beyond ChatGPT. As it ‘explained’ to me:

‘I am not a human being and therefore do not have personal expertise or experience in any particular field. I am a large language model trained by OpenAI to generate human-like text based on the input I receive. I can provide information and answers based on the data I have been trained on, but I do not have personal expertise or knowledge beyond what I have been trained to generate.’

Currently, I think programs like ChatGPT are best suited for solving queries with tight rules and parameters. Queries related to mathematics or coding problems.

The more open-ended the query, the blander the response.

However, it does have ‘creative’ ability. It can generate unique responses when given creative prompts.

Here’s what ChatGPT generated when I asked it to write a ‘funny 250-word script for the TV show Friends’:

‘Ross and Joey are sitting in Central Park, discussing their latest dating exploits.

‘Ross: "I can’t believe I went out with that woman who turned out to be a man."

‘Joey: "Dude, that’s rough. But at least you found out before things got too serious."

‘Ross: "Yeah, I guess that’s true. But now I’m back to square one with my love life."

‘Joey: "Don’t worry, Ross. You’ll find someone. You just have to keep putting yourself out there."

‘Ross: "Easier said than done, Joey. I’m not getting any younger, you know."

‘Joey: "Ah, age is just a number, my friend. You’re still a young, virile man. Women can’t resist you."

‘Ross: "Thanks, Joey. You always know how to make me feel better."

‘Joey: "Of course, Ross. That’s what friends are for. Now let’s go grab some more coffee before we have to head back to work."

‘As they get up to leave, Ross trips and falls, landing face-first on the floor.

‘Joey: "Are you okay, Ross?"

‘Ross: "Yeah, I’m fine. Just a little embarrassed."

‘Joey: "Don’t be, Ross. We’ve all been there. Just brush yourself off and keep on moving."

‘Ross: "Thanks, Joey. You’re a true friend."

‘Joey: "Anytime, Ross. Anytime."

‘The two friends share a laugh and head off to grab their coffee, with Ross still rubbing his sore nose.’

I’ll admit I laughed at how unfunny this was, but I think this is an interesting vista for programs like ChatGPT.

Here’s a poem ChatGPT generated for me on interest rates and the Reserve Bank of Australia in the style of WH Auden:

‘In the land of Australia,

‘The Reserve Bank holds sway,

‘Over interest rates that ebb and flow,

‘Each and every day.

‘A delicate balance they must keep,

‘Between growth and inflation,

‘To steer the economy clear of harm,

‘And avoid a sharp deflation.

‘They raise and lower rates with care,

‘A powerful tool in their hand,

‘But with great power comes great responsibility,

‘To protect the nation’s land.

‘So let us trust in the RBA,

‘To make decisions sound and wise,

‘For the sake of our economy,

‘And the future of our lives.’

Let’s hope when the Reserve Bank members meet this afternoon, they heed the message.

What will 2032 look like for AI?

In the 1960s, Stanford computer scientist Roy Amara said people ‘overestimate the impact of technology in the short-term and underestimate the effect in the long run.’

Bill Gates echoed that statement years later in one of his books, writing that we overestimate what can be achieved in one year but underestimate what can be achieved in 10.

So if this is what ChatGPT can do now, what will it be capable of in 10 years?

The more I interact with AI products like ChatGPT, the more I think the AI future is nearer than we think.

The technology is advancing rapidly.

What is lagging behind is use cases.

I don’t think we have yet solidified what we can do and what we should do with AI. That will take time, experimentation, and mass discussion.

Firms like OpenAI will help in application discovery, too, by unleashing free programs like DALL.E and ChatGPT on a curious public.

ChatGPT itself suggested some potential applications to me:

Many of these suggested use cases are already a reality.

The Wall Street Journal (WSJ) ran a story last month on how much of the content we encounter on the internet is AI-generated.

Echoing ChatGPT’s use cases, the WSJ reported that most of what we read on the internet could soon be autogenerated:

‘By 2025 or 2030, 90% of the content on the internet will be auto-generated, says Nina Schick, author of a 2020 book about generative AI and its pitfalls. It’s not that nine out of every 10 things we see will be auto-generated, but that automatic generation will hugely increase the volume of content available, she adds. Some of this could come in the form of personalization, such as marketing messages containing synthetic video or actors tuned to our individual tastes. In addition, a lot of it could just be auto-generated content shared on social media, like text or video clips people create with no more effort than what’s required to enter a text prompt into a content-generation service.’

The AI future is coming. In some cases, it’s already here.

I’ll leave you with a haiku ChatGPT sent me on the future of AI and the fate of humanity:

‘Humanity and AI

‘Two sides of the same coin

‘Drawn together, bound.’

Until next week,

Kiryll Prakapenka,
For Money Morning