I can bet that many of you have spat this exasperated question at a techie friend in the last six months. Being on a public forum, I’ve anonymised the swear word, but in private conversations, I can equally bet that you would have exclaimed the formidable four-letter word in all its fury!
I can also bet that your exasperation would have grown with the techie friend’s mumbo jumbo response:
See, AI is not intelligence, ie the computer cannot “think” or “understand” like you and me. What AI does is read trillions of word patterns in nanoseconds to predict what the next word will be.
See, layered mathematical functions in a Large Language Model (LLM) can read billions of texts to accurately predict what the next word should be, and it can keep on doing this for thousands of words to produce an output that appears fully reasoned.
See, the computer cannot reason but can mimic reasoning!!
By now you are ready to clobber your see-sawing techie friend.
It’s called intelligence, but it’s not.
It’s cutting-edge science, but it has no powers of reasoning.
Heck, what the *@#$ is AI?
I Don’t Need to Understand AI; I Just Love What it DOES
I’ve given up trying to understand this animal.
I leave that to the nerds.
But I’ve gotten addicted to what AI can do.
It’s *@#$-ing magical!
Earlier, I needed two researchers digging through tomes for about a week to prepare a brief. Now I can do it all by myself on ChatGPT in an hour. Earlier, my engineering colleagues in Bangalore would need six coders over six months to custom-build a website and its CMS (Content Management System). Now one coder, doubling up as an AI prompt engineer, can do the same job in less than a month.
Why, even this article will get diced by Newseasy, an AI editor we’ve created at The Quint. Check it out right here. That stuff in highlighted boxes has been created by this animal.
Exasperating! How does this magician summarise my copy better and faster than I could?
“It’s LLM”, my techie colleagues yell. No no sir, I am not going there again!
AI is Upheaval-ing Our World
If a single AI chatbot has reconfigured my little nest completely, what could it be doing to the big bad world outside? A thousand colourful descriptors spring to mind. It’s inverting, destroying, reconstructing … actually, it’s upheaval-ing (even if that’s not a legit word, I will use it artificially) the 20th century framework that my generation grew up in.
Have you heard of a “dark factory”? It’s literally that, completely dark, without any electric lighting. It’s a factory run entirely by robots, not requiring any (or very negligible) human intervention. The most iconic dark factory is, where else, but in China!
The Xiaomi plant in Changping is sprawled over 81,000 sq meters. It produces 10 million devices every year, that’s one smartphone every three seconds! But there’s not a single human being on the shop floor, so there’s no need for lights or toilets or coffee vending machines. Robots don’t need any of these human amenities. Instead, their “lifeblood” is AI, Industrial Internet of Things (IIOT), machine vision, and cybersecurity.
With such cutting-edge technology running through their metallic veins, these robots “work” incessantly, 24x7, never stopping to gossip or scratch their nose. It took a mere $300 mn (Rs 2,500 cr) to make human labour redundant.
And if you think Xiaomi’s “lights out” plant is an isolated exception, go figure!
The FANUC dark factory in Japan is a plus point in irony. It uses robots to make robots. The shop floor remains dark for 30 days at a stretch, that’s how long these humanoids can continue uninterrupted work. Then lights come on, and a few guys in flesh and blood trudge in for preventive maintenance. Before long, the lights go off, and humans exit. The robots get to work again for 30 continuous days, 24x7. I reckon it won’t be long before robots take on housekeeping functions and the factory goes dark perennially.
Will AI Invert The Classical Capital/Labour Equation?
How fast are robots replacing human labour? There are some indicative stats that capture this phenomenon. More than five million robots are “working” on earth, and over half a million are “newly born” per annum. Put another way, roughly 13 robots are getting added every year for a block of 10,000 human workers. The fastest addition is in Asia, led by China, followed by Europe and America.
As a rule of thumb, each robot displaces about 3.3 jobs, so these machines could be killing 43 jobs per 10,000 workers or replacing about 0.43 percent of human jobs every year. That does not sound alarming, but it is.
For one, the trend is accelerating. For another, these robots are getting more efficient and productive, so the “homicide rate” could be swiftly moving up, ie each robot could end up killing many more than just 3.3 jobs with every passing day. Now that AI is powering robotics, there could be a chilling spree of “mass murders”, with human labour getting replaced exponentially in automotive, electronics, metals, and machinery verticals.
Classical economic theory has taught us that capital and labour are two critical factors of production. Their relative value contribution remained remarkably constant through the 20th century, with capital accounting for about 30-40 per cent, and labour picking up 60-70 percent of output.
But the equation is about to invert in the 21st century. Robots are machines, ie they are capital, but they perform labour’s role by replacing human beings, ie they are diminishing labour’s relative contribution to output.
We’ve already approximated, above, that robots, or capital, could be replacing labour at the rate of about half a percentage point per annum. So, the capital/labour ratio is rapidly toppling in favour of capital - in a few cataclysmic years, capital could pick up 60-70 percent of income/output, while labour could struggle to retain its share at 30-40 percent.
What happens then? When capital captures the bulk of global income, and labour is stranded with menial or no jobs?
It’s an existential question for humankind, with uncharted consequences.
We shall probe this stunning inversion in Part 2 of our enquiry into a post-AI world.