The horseshit paradox: Why fears about tech are wildly exaggerated
I have believed for a long time that the rise of artificial intelligence (AI) could pose a threat to our notions of work. Would it enable us to spend more of our time doing what we love? I am beginning to believe so, particularly after a recent conversation with a friend.
She is returning to college to pursue a degree in education and she sounded optimistic about the future. This despite the fact that she and her family will have to significantly alter their lifestyle when she, now in her early 40s, logs out of her full-time job in the travel and tourism sector. But the pandemic has already hit that sector hard. She knows her “technological displacement” is imminent, she says, and must reinvent herself.
Her optimism is echoed by the economist Daniel Susskind in his book A World Without Work. In the 1890s, he writes, citizens of New York and London were witness to hundreds of thousands of horses heaving all kinds of vehicles through their cities. While the horses did their job, they dumped tonnes of manure on the streets.
There were estimates that by the middle of the 19th century, the streets of New York City would be nine feet deep in horseshit, if matters continued as they were. Banning horses was not an option; the economy would collapse. In policymaking circles, this was referred to as the “Horseshit Paradox”.
While thought leaders of the time were fretting, they did not notice the emergence of the internal combustion engine. It could create cars. When those cars hit the roads, there was an outcry. What rules ought cars to follow? How many horse stations and caretakers would be displaced?
By 1912, there were more cars on the streets than horses. Five years later, the last horse-drawn tram was decommissioned from service in New York. Few mourned its demise. The city had entered the age of the automobile.
The Horseshit Paradox was held up to describe the upside of technology until the 1980s. Then a Russian economist Wassily Leontief put into words what so many already feared — that what cars had done to horses, computers could do to humans. The first time this became real for me — the first time I felt anxious about the future of my career and that of my children — was when chess grandmaster Garry Kasparov was beaten by the computer programme Deep Blue in 1997.
If an algorithm could outthink Kasparov, how long before I was toast?
By 2012, companies such as Narrative Science were designing algorithms that could generate news reports more quickly than a reporter (albeit with no scepticism and little nuance). Algorithms are working fast and smart in the fields of financial services, medicine, law and have now begun to “create art” as well. Perhaps Leontief was right?
On reading Susskind closely, though, my concerns would appear to be vastly exaggerated. The reason a computer could beat Kasparov was because it could compute as many as 100 billion to 200 billion moves in three minutes. It was brute force. When Kasparov reflected upon the game later, in his lovely book Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins, he wrote: “Chess just isn’t complex enough and even I can admit that there is more to life than chess.”
There is also more to Kasparov than chess. And there is more to the reporter than her news reports. Each of these cutting-edge computer programmes can, at least for now, only do one task, while stationary, with strict limits to the complexity they can handle within that task.
That’s not how humans function. As Kasparov pointed out in his book: “Applying context comes naturally to humans… Our brain does the work in the background without any noticeable effort on our part, nearly as effortlessly as breathing.”
Simple tasks completed quickly would not make the world go round. And so I think, yes, we must future-proof ourselves. Expand the horizons of what we know, do and learn about. But all the while acknowledge that there is no mechanical substitute, as of now, for the human brain.