By Brendan McDermid
See original post here.
Krugman, who regularly writes about economic policies and their broader impact, said he thinks AI is “too pervasive and too diffuse” to warrant any major intervention from the government.
What he means is that the AI boom is likely to wipe out some white-collar jobs (over manual-labor jobs) evenly across the country — instead of devastating one particular region or social group. That means it’ll be hard to design remedial policies, he says, to help displaced workers.
“I don’t think there’s much you can do about that aside from just trying to ban the technology altogether, which isn’t going to work,” Krugman said.
Krugman contended that a lot of the white-collar work that tools like ChatGPT are threatening, though, isn’t rooted in creative or original thinking.
“How many people out there in the real world are in fact being creative? How much of the work that we pay people a lot of money to do is in fact a lot like super-expanded autocorrect? And I think the answer is quite a lot. So this is potentially a really big thing, and it could displace a lot of jobs,” he said.
But how does “a really big thing” translate into a measurable impact on the economy? Krugman said it’s still hard to gauge the impact of technological innovation like AI because the way we “measure technology is godawful.”
Krugman explained that economists usually add up tangible metrics like “increased stock of capital to economic growth” and chalk up what can’t be measured to technological growth — because it’s not clear how else to measure it.
For example, he said the country’s productivity accelerated at a rate of 1% a year between the mid-90s and early 2000s, which has been associated with the advent of the internet and companies finally figuring out how to leverage information technology.
Some of the growth, though, also just came from more commonplace things like companies figuring out how to use barcodes to manage their inventories, he added.
“It’s basically technology is the measure of what you can’t explain otherwise,” Krugman said. What it will mean when it comes to AI, only time will tell.