The famous opening scene of Stanley Kubrick’s 2001: A Space Odyssey, shows a group of apes jostling for position. One ape discovers a bone, a femur, and begins to use it as a tool to bash things. This primate, armed with the tool, emerges as the top dog. Triumphant, he throws the white bone in the air, which Kubrick miraculously transforms into a gleaming white spaceship circumnavigating the planet. Stanley Kubrick, in one scene, captured what economists have been going on about for years. Technology enables progress, whether it be the early primates armed with a bone, modern humans equipped with space travelling engines or ChatGPT, perhaps the most impressive technology I’ve encountered.
Typically late to the party, I tried out, for the first time this week, the miracle of ChatGPT. Sometimes there’s a moment when a technology transforms our world. It may have been when you saw the first smartphone in the 2000s or when you first experienced the internet in the 1990s, but those moments stay with us. Last weekend, struggling to write briefs for 60 events at Kilkenomics, an economics and comedy festival in Kilkenny, a colleague suggested I type the topics and question angles into ChatGPT. Almost instantaneously, up came a list of detailed questions, with some fascinating prompts, including references, which looked accurate. It wasn’t perfect (in my eyes) but it was almost there and, as the technology develops, it will get better and better.
Scraping the internet for everything that has ever been uploaded on a subject, ranking the material and collating it in a structured and intelligent way, the possibilities of what are called “large language models” are mind-blowing. Think of all the amazing new jobs that will be created, all the creativity that will be released by absolving people of the drudgery of mind-numbingly repetitive work such as creating economic models of the economy! And that’s just in my field. Imagine other fields! For a services economy such as Ireland, where thousands are involved in clerical jobs, the liberation from drudge will be immense.
Like every new technology, there are warnings that this new tool will put thousands out of work, permanently. It is natural to be uneasy when faced with something that can do your job better than you – but new technology has been a significant contributor to what we call human development. Humans have been problem solvers from the very beginning, as Stanley Kubrick reminded us. Every technological innovation allows us to solve a problem, move on and solve the next one.
The interesting social aspect of this technology is that it will affect the middle classes and white-collar jobs most. In the past, technological innovation was primarily used to automate working-class jobs
The pace at which ChatGPT is being adopted is mesmerising, suggesting it is solving a very big problem or something that many people believe is a big problem.It notched up 100 million monthly users within two months of launching. To put this pace of growth in context, it took Facebook over four years to hit that number, five years for Twitter and over two years for Instagram. Now the company behind it claims that 100 million people are using ChatGPT on a weekly basis, making it the fastest-growing app of all time. It is growing and adapting all the time. Until a few weeks ago, only information on the internet up to January 2022 was available to its algorithm. As of today, all events up to April 2023 are covered. The size of the underlying models and their speed is phenomenal.
The interesting social aspect of this technology is that it will affect the middle classes and white-collar jobs most. In the past, technological innovation was primarily used to automate working-class jobs. In farming, we saw tractors eliminate the need for low-paid labourers, while on the factory floor robots replaced human workers. Many of the so-called “left behinds” who voted for populist parties in recent years live in former industrial areas where automation replaced manufacturing jobs. The factories moved or shut down but the people who worked in these jobs remained – in place but unmoored.
However, these specific instances mask the fact that far more people availed of technology than were made redundant by it. In rural Ireland of the 1960s and 1970s, many warned that the arrival of the tractor would destroy farm labourer jobs. This did happen and today Ireland has its fewest number of farmers working the land – but the greatest agricultural output it has ever had. Those people who left the land are now in the workforce, a workforce that despite – or maybe because of – constant technological advances, is the largest this country has ever experienced.
[ Can AI be taught to stop straying from its ethical guardrails?Opens in new window ]
What happens when the middle class get hit by ChatGPT? Consider the well-paying, high-end clerical jobs such as solicitors, financial back-office reporting, accounting or consultancy. Think about the report writers, the journalists, freelancers, graphic designers, and civil service departments churning out papers; what happens to them? On one level, many rudimentary functions that are clogging up people’s days will be done by this piece of technology, giving people time to do other things. But many white-collar professionals are creatures of habit, reasonably well off and happy in their jobs. What will happen if a machine can do their job at a fraction of the cost? Will they be able to deal with the shock of redundancy?
The economy is always switched on, always in flux, never at equilibrium. The problem is that sometimes we humans just want to slow down, grab a breath or yearn to opt out
A recent article in the Financial Times emphasises that highly skilled workers are the ones who will face the detrimental consequences of AI first. Specifically, “those with a six-figure salary are about three times as exposed [to risk] as someone making $30,000″. Citing a paper that focuses on the effects of generative AI since November 2022 (when OpenAI released ChatGPT to the public), the conclusion is that freelancers were among the most affected. In certain fields, they experienced a decrease of 2 per cent in the number of monthly jobs and a decrease of 5.2 per cent in monthly earnings.
[ Artificial intelligence could pose ‘major threat’ to college qualificationsOpens in new window ]
Think now about the next few years for copywriters, advertising creatives, university lecturers, economic researchers and the report-writing industry in general. Of course, we will adapt. The history of economics is the history of technological innovation and adaptation followed by yet more innovation and more adaptation. The economy is always switched on, always in flux, never at equilibrium. The problem is that sometimes we humans just want to slow down, grab a breath or yearn to opt out.
We are entering a new era. This will be as big as the printing press, the industrial revolution and the internet age put together. Welcome to the world of large language models. A new human and quasi-human Odyssey. With no stop button.