When are we all going to lose our jobs to AI?

4 Mar 2024

Sweden’s Klarna announced that they have replaced 700 mid-level customer service workers with a customised AI-chatbot. This, of course, is a smidgeon of a much larger story. What is going to happen to all of us as AI determinedly marches its way across the landscape of human capabilities?

An eyebrow-raising item of news came out of Sweden a few days ago, and very quickly went viral. That gentle social democratic country is home to a payments and shopping services giant called Klarna. They serve 150 million customers and 50,000 merchants worldwide.

Klarna announced that they have replaced 700 mid-level customer service workers with a customised AI-chatbot.

Adding insult to injury (at least to those who have lost their jobs) were some of the background stats. The Chatbot speaks 35 languages. It results in 25% fewer repeat calls. It works 24 hours a day. It reports a customer satisfaction rating on a par with human agents. It handled 2.3 million calls last month, which was 75% of all calls to Klarna. It is dramatically less expensive than its human equivalents. And the company is now set to expand AI into all corners of its business.

This got me wondering which skills and jobs are going to cede ground to AI, and when.

It goes without saying that today the chatbot is the worst it is ever going to be. It will be better tomorrow. It will continually improve — pretty much on its own, which is machine learning’s scary superpower.

A much larger story

This, of course, is a smidgeon of a much larger story — that of AI and employment, and what is going to happen to all of us as AI determinedly marches its way across the landscape of human capabilities.

Take law, for instance. I was recently in an upmarket bar having a beer with a highly regarded tech-savvy lawyer. Discussion moved to the intrusion of AI into the legal profession. He pulled out his iPhone and spoke to it. “What is the law governing the distribution of company dividends under (certain conditions)? And please quote the related legal precedents.” (His query was obviously much more specific than this — he named the “certain conditions” in appropriately complex legal jargon.) His iPhone waited a beat and then responded clearly and succinctly to the request with a crisp 60 second audio response — which it was prepared to deepen upon request.

Read more in Daily Maverick: And the AI said: Let there be Light! And there was Light…

My lawyer colleague’s company has developed this bot, and his pride was palpable. The research that would have been required to answer his query would previously have taken many hours, perhaps days. The bot took seconds. We did not discuss what happens with regard to the billable hours saved with this coming wave, or who benefits from it, but it was clear that much of the work of lawyers, up to and including the formulation of complex and deeply reasoned positions and opinions (a very high-level skill) will quickly be wrested from its comfortable position in the legal profession (at least partially, if not wholly). Some lawyers will be thrilled and welcome this with open arms. Many, I suspect, will feel differently.

Of course, some of you will remember an embarrassing news item from a few months back (in the dark ages of AI) which reported that a lawyer was punished for presenting an argument that included some precedents excavated from ChatGPT, which turned out to have been completely made up by the ghost in the machine. Most people concluded that AI was a long way from prime time, at least within a highly skilled profession like law.

Not anymore. The technology demonstrated to me in the bar belongs to the new breed of “GPTs” (not to be confused with ChatGPT). These GPTs are highly specialised and focussed on specific fields, trained on a limited body of content. This specialisation results in their being very much more accurate in their field than the big, generalised language models that blanket the headlines.

AI, skills and jobs

This got me wondering which skills and jobs are going to cede ground to AI, and when. The answer arrived in the form of a startling new academic survey of 2,788 AI professionals who were polled on this very topic. It is titled very literally ‘Thousands of AI Authors on the Future of AI’. It was released in January of this year.

Here is a paragraph taken from the abstract:

“If science continues undisrupted, the chance of unaided machines outperforming humans in every possible task was estimated at 10% by 2027, and 50% by 2047. The latter estimate is 13 years earlier than that reached in a similar survey we conducted only one year earlier [Grace et al., 2022]. However, the chance of all human occupations becoming fully automatable was forecast to reach 10% by 2037, and 50% as late as 2116 (compared to 2164 in the 2022 survey).”

I have italicised the second sentence because, in less than two years, researchers have reduced their estimate by 13 years! Meaning (obviously) that AI-capability is accelerating way faster than experts imagined a very short while ago.

And then there is the third sentence, perhaps even more shocking. By 2116, long after we are all gone, but perhaps within the lifetime of some of our children, there is a 50% chance that all occupations will be fully automatable. The study is replete with charts and graphs outlining the chances of employment carnage.

Think about it. Lawyers, surgeons, chefs, CEOs, janitors, teachers, writers, musicians, engineers, programmers, architects, gardeners, accountants, short-order cooks, pilots, interior designers, manicurists, dentists, mechanics, bin pickers, electricians, plumbers…

Read more in Daily Maverick: AI shaping up to become the greatest geopolitical weapon in history

All of them, including those jobs that require physical dexterity (this is one of AI’s next big horizons). What’s more, I am confident that, if a new survey is conducted in two years’ time, it will bring that horizon even closer, because humans are really bad at extrapolating exponential growth.

There are those who say that this is all nonsense, that tech has always caused job migration and the requirement to upskill, but it has also been the major contributor to ever-increasing employment throughout history. Indeed, there is much data to support this.

Perhaps this time is different. We would all like to think not. But, if you are one of Klarna’s 700 laid-off workers, you might have a different view.

Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg. His new book It’s Mine: How the Crypto Industry is Redefining Ownership is published by Maverick451 in SA and Legend Times Group in UK/EU, available now.

Categories:

Tags:

DM is much better if you are signed in...

Business Maverick

A much larger story

AI, skills and jobs