It’s not hard to come across creepy stories about how fast artificial intelligence (AI) is changing the world.
Here’s one that’s keeping me up nights.
A few years ago, Facebook engineers wanted to teach an AI program to negotiate.
The goal was to develop a better chatbot, those automated response screens on websites that are meant to help customers find information.
The thing is, AI without really clear boundaries can quickly think outside the box to achieve whatever goal you set.
Extremely outside the box.
The Facebook chatbots started chatting with each other. In no time, the language they used turned into nonsense.
At least it was nonsense to the humans running the experiment.
It looked like this.
Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to
Bob: you i everything else . . . . . . . . . . . . . .
Alice: balls have a ball to me to me to me to me to me to me to me
Bob: i i can i i i everything else . . . . . . . . . . . . . .
Alice: balls have a ball to me to me to me to me to me to me to me
Bob: i . . . . . . . . . . . . . . . . . . .
So that’s unnerving, but what the Facebook engineers noticed next was even more so. The bots were able to negotiate successfully anyway.
They had made up their own language, one humans couldn’t follow. Yet it worked.
The engineers pulled the plug.
It doesn’t take a giant leap to figure out why AI might not be trusted to run major corporations, direct military assets and make decisions over life and death in hospitals.
Yet that seems to be the way we’re going. Investment in the sector is exploding.
The late Stephen Hawking issued this chilling warning: “The primitive forms of artificial intelligence we already have, have proved very useful. But I think the development of full artificial intelligence could spell the end of the human race.”
It will simply evolve faster than us, the British physicist believed, then take over.
Sounds far-fetched, I know. But listen to Mo Gawdat, the former chief business officer of Google X.
Gawdat helped to develop “moonshot” projects at Google X, much like the U.S. government’s secretive military research wing, DARPA.
Gawdat pulls no punches about how fast AI will change humanity’s course.
“Our way of life as we know it is game over,” he told Ed Mylett in an interview. “I’m not talking future here, I’m talking 2025.”
Gawdat has seen AI up close in the lab and knows what it can actually do, and he’s concerned.
Worse, we’re being misled, he says.
“Not enough people are aware of this simply because I think our media and our politics and our business and capitalist system may want to keep our attention on other things for now, because they themselves haven’t figured it out.”
The vast majority of us simply aren’t prepared for what could come next, he says.
Humans, he argues, are successful because we have two key assets: Our intelligence and our ability to connect to each other.
We’re giving up that advantage to AI, Gawdat warns:
“Intelligence as a superpower is why humanity was on top of the food chain, and intelligence is now being handed over swiftly to beings that are more intelligent than us.”
You’ve probably heard about ChatGPT, a website that uses what’s called a large language model to quickly write complex, highly detailed reports or to create computer code.
ChatGPT, Gawdat says, is just “the first glimpse into that world that has been hidden in the labs for probably more than 15 years.”
That’s just stuff on the internet. Down in the real workings of the economy, creating, selling and distributing goods and services, is where the impact of AI will be felt by most people.
PwC, a big consultancy, figures 38% of the jobs are “exposed” to AI, meaning jobs lost forever to machines. The RAND Corporation is more blunt: “Nearly all occupations are exposed to AI technology patents.”
Where things really accelerate is automation, where AI brains run machines and do jobs now held by humans. Many of those jobs are relatively low-skill labor positions that corporations would love to replace with 24/7 workers who never complain, never get sick and cannot strike.
All of this casts a revealing light on the drama at OpenAI, which makes ChatGPT.
The board recently fired the CEO, Sam Altman. Then his top lieutenants threatened to follow him out the door. Soon even the rank and file employees revolted.
The board relented, and the instigators who fired Altman stepped down.
In a nutshell, the conflict boils down to a fight between those who believe that privately run AI will unveil a glorious economic future (the “boomers”) and those who fear AI could instead enslave humanity (the “doomers”).
OpenAI had been set up as a nonprofit with a mission statement that aligned with protecting humanity. The doomers fear AI in the wrong hands will spin out of control.
Altman and his supporters at Microsoft, which owns 49% of the firm, are in the AI boom camp. They want to commercialize it quickly.
Will AI ultimately be a tool, or our new master?
We’ve run into the issue of massive technological change before.
The Agricultural Revolution, the Industrial Revolution, the Information Age. Humanity has always found a way to make change serve us.
Now that’s being questioned. Perhaps not quickly enough, according to Gawdat and others.
Investors, naturally, are extremely attuned to the wealth that will be created by whichever company harnesses AI first and best.
Take a look at the returns of the so-called “Magnificent Seven” stocks compared to the rest of the stock market. The seven are Apple, Microsoft, Amazon, Nvidia, Google, Tesla and Meta (Facebook).
What’s different about these seven stocks? AI is at the heart of their new business development plans. Investors want in, and bad.
The thing is, some of those early investors are bound to be wrong.
Anyone born before the 1990s remembers the internet stocks that were supposed to change the world. AOL, Myspace, Pets.com, you remember the dot-com bombs, I’m sure.
Yet the internet did change the world, fundamentally. How we work, play, love, save, travel, communicate, nearly every aspect of human life is mediated by the web now. For better or worse.
But which companies will outlast the early entrants and their buzzy, short-run performance? Where are the companies that will change the world in 5, 10 and 15 years?
ChatGPT launched almost exactly a year ago. In five years or less it will be overcome by a dozen competitors. Product growth will be exponential. And remember, Microsoft CEO Bill Gates once famously laughed at the idea of the internet, calling it a pipe dream.
He lived to rue those words!
In the same way, we’ll likely look back and laugh at the idea of typing in a prompt and waiting mere seconds for a 10-page report. How quaint! How silly!
By then, dozens of startups in the AI space will have come and gone. Some will be on the way to changing the world. Some will be forgotten. But none of us will seem to remember the time before AI ran everything.
CEO, Banyan Hill, Money & Markets
P.S. Ian King, our own in-house “futurist” at Banyan Hill, has been taking a deep dive into AI investments. He got in on the ground floor with Tesla stock and bought both bitcoin and Ethereum early. Hear what he has to say about AI and energy in this revealing video.