The future is finally here

Artificial intelligence has the potential to not only reduce costs but change the architecture of the office entirely

In 1987 the economist and Nobel prize-winner Robert Solow gave a talk in which he questioned the economic value of the personal computer. ‘Why’, he asked, ‘do we see computers everywhere but in productivity statistics?’

What he meant was that the microchip, the core innovation of the personal computer, had already existed several decades earlier. One of the most popular personal computers, the Altair 8800, was released in 1975. Many economists predicted spectacular productivity increases and yet, by the end of the 1980s, there was no evidence in the productivity statistics that computers had made any difference.

We know, of course, what happened after Solow’s talk. Personal computers, combined with the internet, would radically change the office and the type of work we do. In chapter 31 of Our Long Walk to Economic Freedom (that will be published in Afrikaans next week, more on that next week), I tell the story of Paul David, the economic historian who was to explain Solow’s paradox. David used a historical example to explain the phenomenon: the electric dynamo’s invention in the 19th century was just as transformational for society, but equally took several decades before its impact was noted. This is because the dynamo was a general-purpose technology, one that affected not only the cost of production but the entire structure of production.

What was true for personal computers in the 1990s (and the dynamo in the 19th century) is true for artificial intelligence today. Excitement about artificial intelligence and its applications was already a popular topic in mainstream media by the late 2000s. While AI drew much earlier attention from scientists, advances in deep learning techniques only brought its usefulness to the public’s attention with the development of image and speech recognition services, and games. In 2011 IBM’s Watson computer beat the Jeopardy!-champion, which catapulted AI to the front pages of newspapers. Around that time, Google began investing heavily in deep learning R&D and released services like Google Translate and Google Photos. Amazon’s Alexa and Apple’s Siri would follow soon. Other applications, like Tesla’s self-driving cars and pharmaceutical companies’ ability to produce vaccines quickly during Covid, showed the general applicability of the technology. But, in truth, most of us would have been able to live and work without using AI.

The arrival of ChatGPT has changed that. I asked ChatGPT how it would explain itself to a ten-year-old: ‘ChatGPT is a computer program that can talk to people and answer their questions like a really smart robot.’ That seems self-explanatory enough. But why is this so revolutionary?

In short: AI platforms like ChatGPT have the potential to radically alter the tasks most of us perform on a daily basis. To name just one example: business. Two MIT economists, Shakked Noy and Whitney Zhang, show in a recent paper that ChapGPT can increase the productivity of university-educated workers by more than 50%. They recruited 444 workers in an experiment, gave half of them access to ChatGPT, and then asked everyone to perform various tasks. Not only were those with ChatGPT more productive, but AI benefited the typically least-productive workers, closing the gap between the best and slowest.

ChatGPT is just one of many AI platforms we’ll see soon. Microsoft’s Bing is already available and more powerful. Although I’ve not been able to gain access yet, several scholars have been able to test its ability. Ethan Mollick of the Wharton School at the University of Pennsylvania is one of them. He writes: ‘I find the abilities of AI to be a constant source of amazement, and … think a focus on just using them for writing and personal assistant tasks ignores some of the most exciting things about AI.’

So what is it capable of? In one test, Mollick asks Bing to ‘come up with movie ideas for a movie starring the cast of Fast and the Furious 8, except The Rock is replaced with a puppet of a dolphin. The movie is a historical drama set in the 1300s, and also incorporates at least one William Carlos William poem into the plot. The movie should also act as commentary on a youth fad of the 20th century.’

The response can be found on his Substack account, but it is remarkable. As I’ve noted before, these AI platforms are not creative in the typical sense of producing something novel, but they are immensely powerful synthesisers, which means they can take these disparate topics and fuse them into something that seems novel.

But what are the practical implications of a story of a puppet dolphin? Well, endless, to be honest.

As soon as you realise the power of this technology, it changes how you perform almost every task you’re busy with. You’ve written something that must be shortened by half (or doubled in length)? Ask the AI to do it. Need to complete a report, or an application, or any paperwork, really? Ask the AI to do it. Need to do a presentation about new marketing channels, product ideas, or cost-cutting strategies? Ask the AI.

In Power and Prediction: The Disruptive Economics of Artificial Intelligence, the economists Ajay Agrawal, Joshua Gans and Avi Goldfarb compares artificial intelligence to the arrival of electricity. Not only did the price of power fall dramatically, but electricity allowed (or, actually, forced) a change in the factory's architecture, creating efficiencies that would not have been possible with other power sources. In other words, not only are the same tasks cheaper than before, but because of the system’s redesign, we can do entirely new things with it.

Artificial intelligence is now doing the same. Who will benefit? Probably smaller firms which have smaller legacy infrastructures and can adapt quicker. And who will lose? Well, certain types of jobs will certainly require big readjustments. Language editors, for example, won’t sleep comfortably. (Grammarly, for example, has already improved my writing to the extent that I don’t need a language editor for most things.) But for most workers, AI will be complementary rather than a substitute. ChatGPT will remove the mundane tasks: report writing, form filling, etc., allowing us to focus on the things of higher value which require human interaction.

These things will help to cut costs. But more than that, the real efficiency lies in the transformation of the entire architecture of the office in ways that we probably cannot yet imagine. By combining artificial intelligence and remote work, the office of the future might look very different than today. One possibility is that this might help alleviate the huge worker shortage in the rich world or at least make possible the outsourcing of certain tasks that can be performed better by machines or people in other countries. Another possibility is the displacement of the clinical office environment with spaces that allow for team activities and leisure. Instead of an eight-to-five office job, perhaps future workers will only meet for an extended and informal lunch three days a week.

Who knows? But what is guaranteed is that artificial intelligence is finally bringing the future much closer, making life easier, more fun and more meaningful. And that can only be a good thing.

An edited version of this post appeared (in Afrikaans) in Rapport on 12 March 2023. Photo by Shubham Dhage on Unsplash.