If Friday’s jobs numbers are anywhere near as spectacular as last month’s, maybe no one will care if ChatGPT and its brace of new competitors begin doing more of the brain work that used to be reserved for humans.
Sam Altman, chief executive of OpenAI, the company that brought the new intelligent software tool ChatGPT to the public last November, recently declared that artificial intelligence (AI) could eventually “break capitalism.”
But many experienced AI researchers are worried that in the shorter term it could shift the balance of power in favour of capital, increasing inequality.
In a world where we have become accustomed to shareholders owning the factories — and more recently, the robots — that make the things we need, Canadians who keep a close eye on artificial intelligence worry that the revolutionary nature of the new technology could upset the world of work in a way we have not seen with previous innovations.
Who owns intelligence?
In his latest book Power and Prediction, The Disruptive Economics of Artificial Intelligence, Canadian AI authority and entrepreneur Ajay Agrawal reassures us that AI continues to be controlled by humans.
That may change if people like Altman succeed in their stated goal of creating artificial general intelligence or AGI, the true thinking machines of science fiction utopia and horror.
But well before the arrival of Arnold Schwarzenegger’s Terminator robot and its boss Skynet, the more practical question that concerns those who study our relationship with computerized systems is: Which humans are the ones that will control the AI?
Certainly the world of money has been rushing to get a piece of the action. In a phone conversation this week, Agrawal said executives from industries including banking, insurance, pharmaceuticals and car-making are pestering him to ask how it will affect them. Just this week Google announced it was investing in its own ChatGPT equivalent, called Bard.
When the Chinese tech giant Baidu announced it was launching a version of its own, called Ernie Bot, its share price jumped to an 11-month high, despite the goofy name. Microsoft says it will parlay its stake in ChatGPT to transform its also-ran search engine Bing and give Google a run for its money.
While early play with ChatGPT has been free, helping it go viral, OpenAI announced last week it would soon roll out a premium version for $20 US per month — setting the stage for different tiers of access to any advantages it brings.
Like children released on a jungle gym
In some ways the surprising thing about ChatGPT is how it caught not just the general public, but even artificial intelligence experts by surprise.
People like Karina Vold, a philosopher of cognitive science and artificial intelligence at the University of Toronto, knew this kind of thing was around the corner, but the user-friendly accessibility that allowed almost anyone with a few computer skills to try it out has been transformative. She thinks even its creators were surprised.
“They are learning, I think, a lot from our own human feedback as we play with the system, kind of like building a jungle gym and then releasing a bunch of children onto it,” said Vold.
She said that as millions of us try the software tools, unwittingly training them for new uses, usually after signing implied contracts to give away our rights, human workers are being co-opted for profit without compensation.
“I think both our courts and our policymakers are behind the curve in thinking about how best to manage and regulate these companies that are now larger than many countries,” said Vold.
Information technology specialist Gerald Grant at Carleton University’s Sprott School of Business in Ottawa is among those who say artificial intelligence needs to be regulated before it gets out of hand. History shows that new technologies have the power to disenfranchise some groups in society and the ownership of brain-power may lead to a transfer of wealth from labour to business.
“We need to have a thorough look at what the implications are for these technologies and what it means for the transfer of wealth and capital,” said Grant. But as we’ve seen with Canadian attempts to control online media, calls for regulation are not easy to fulfil.
Instructions in plain English
Joel Blit, an engineer and economist who teaches at Ontario’s University of Waterloo and is helping to prepare a free online seminar later this month titled ChatGPT: What executives really need to know, says it is no surprise that business sees an opportunity in the new software tools.
Like the others I spoke to, he said the sudden appearance of such a user-friendly interface, not just able to write and paint original material but, as Agrawal suggests, to do such things as “task planning” for a physical robot — allowing it to follow instructions given in plain language — represents an inflection point that forces reconsideration of our economic relationships.
“All this capital that has been flowing in there, the algorithms have obviously improved, and I think we are probably getting to a tipping point where things are really going to explode,” said Blit.
In general, people like Blit are what some call “techo-optimists” who see huge advantages for society as we use the power of artificial intelligence to improve the world.
But Blit said there are also “scary” implications: the technology has the potential to suddenly wipe out whole categories of human work at a level that factory machines, computers and robots did not. More than ever, companies will be looking for new ways to use the powerful tool to replace expensive humans with cheap software.
ChatGPT is artificial intelligence chatbot software capable of writing poems, college-level essays and even computer code. Experts say the software highlights how far AI has come in just a few years, while still spotlighting concerns around accuracy.
Like many other optimists, he said it is reasonable that in time humans will find a way to make a living doing things that humans do best, but he worries that may take a long time.
“I think the loss of jobs is real, but to tell you the truth, what I’m most worried about is not the loss of jobs but the potential impact it can have on inequality,” said Blit.
Since artificial intelligence is owned capital, there is a possibility that a larger share of intellectual productivity could go to the owners. But Blit said that even among different types of workers there will be winners and losers, those who are able to use the power of the new technology to improve their income and those who don’t have the proper skills.
Who gets the benefit?
Agrawal offers two contrasting examples, one of a gardener who has trouble communicating with clients because of poor writing skills and a highly skilled doctor.
“This person was a good landscaper but a poor communicator by email,” said Agrawal. “What this tool did was it turned his very broken English email communication with his clients into beautifully composed email.”
In that example, he said, the software will equalize incomes because it takes away the advantages of those who are both good landscapers and good composers of email. He said that many new Canadians with broken English could benefit.
The second example was a doctor who had to write letters convincing insurers to cover his patients’ treatment, a task he was capable of doing, but was tedious and took valuable time.
“That’s going to increase income inequality because it is going to take people at the high end of the income distribution that are already making a lot of money and makes them even more productive,” he said.
Agrawal said he expects that as this powerful technology, known as generative AI, becomes more accurate and is deployed across every industry similar to the way the internet was before it, it will help those who are able to use it and hurt others.
“It’s going to shift wealth to the owners of this capital like any new innovation,” said Agrawal. For some applications we might prefer that its a public utility, he said.
As with any business as more participants enter the field, competition will likely reduce any price advantage. He points to internet search, that despite the huge cost of invention, we get largely for free. And as with previous technologies including the internet and search, generative AI has the potential to make at least some of us more productive.
“Whether in the end it will increase or decrease income equality nobody knows,” said Agrawal. And whether it will increase the wages of workers relative to capital, or the opposite? “I think it’s too early to say.”