Is ChatGPT the answer to Homer Simpson’s prayers?

Imagine a person whose desire for the easy life is stronger than his sense of ethics. And imagine this person gets hold of a cutting-edge computer app that can provide quick answers to tough questions. Then imagine that person is asked a difficult question. Instead of answering it himself, he enters it into the computer, then relaxes for a while. Eventually he delivers the computer response and takes credit for many hours of hard work, none of which he has done.

That’s a pretty good description of a 12-year-old schoolboy I know, who typed a homework question into ChatGPT, played games on his Xbox all evening, and then handed computer work over to the teacher, who gave reviews enthusiastic. “Outstanding effort,” was the teacher’s comment, which, come to think of it, is true.

It’s also a good description of how at least one accountant behaved in response to one of the first digital spreadsheet programs, circa 1980. As reported by Steven Levy in his 1984 Wired article “A Spreadsheet Way of Knowledge,” this accountant, when given “an urgent assignment, sat down with his micro and his spreadsheet, finished it in an hour or two, and left it on his desk for two days. Then he Fed Ex it to the customer and got all kinds of accolades for working overtime.

In its ability to generate plausible answers to a huge variety of questions, ChatGPT is unprecedented. But it has a very clear track record in other ways, from cutting frame to spreadsheet to sat nav. Those precedents give us some clues about what might happen next.

The first insight is that if the technology works well enough, it can be adopted quickly. I’ve often written about how it took more than three decades for the electric motor to catch on. A tremendous amount of rethinking, retraining, and refurbishment was needed before factory owners could reap its benefits.

But not all technologies require such epic transformations. The digital spreadsheet broke into the business world in about five years. It was just too good and too easy to use compared to the handwritten alternatives.

Secondly, new technologies do not necessarily destroy jobs, even in the sectors most directly affected. The Planet Money podcast calculated that between 1980 (around when digital spreadsheets first began to be used commercially) and 2015, the US accounting profession lost 400,000 jobs and lost earned 600,000. The lost jobs were often accounting clerks, whose role was to grind the arithmetic through calculators. The jobs that were earned were mostly – dare we say? — creative accountants.

But it’s the third intuition that most intrigues me: different technologies tilt the playing field in different directions. The spreadsheet multiplies the skills of an expert user, but the sat nav is different; it is an alternative to expertise.

The shearing loom disrupted the lives of skilled textile workers because it made a difficult and highly skilled job within the reach of almost everyone. Its use was despised by the Luddite rebels because, like the satellite navigator, it made their expertise superfluous.

The digital spreadsheet is an example of “skills-driven technology change” that helps productive people be even more productive. For nearly half a century, skill-based technological change has been the norm and a major reason why income inequality has increased over the decades. But as the sat nav and the cutting loom demonstrate, some new technologies improve the productivity of less skilled workers. That won’t automatically reduce inequality: Looms may have helped unskilled workers a little, but they mostly benefited capitalists.

What about generative AI systems like ChatGPT and Bard? Do they multiply the output of the elite workers or provide the most help to those who need it? It’s too early to be sure, but the early evidence is intriguing.

A study, by economists Erik Brynjolfsson, Danielle Li and Lindsey Raymond, examined what happened when an AI-powered conversational assistant was deployed to a workforce of more than 5,000 customer service agents working for a software company. These workers typically engage in long text chats with frustrated customers, trying to resolve technical issues. Meanwhile, the chatbot would analyze the chat and suggest possible responses for the customer service representative, which they could use, ignore or adapt.

Brynjolfsson and his colleagues found that chatbots helped: Workers solved slightly more problems than their customers, and they did it 14% faster. And the chatbots weren’t compromised by skills: The best and most experienced agents didn’t benefit from the chatbot, while the least experienced and skilled workers resolved 35% more queries per hour. Those inexperienced workers also learned and improved faster than those without chatbot access.

Another study, by economists Rocked Noy and Whitney Zhang, gave people writing assignments. Half of them had access to ChatGPT, the other half didn’t. Again, it was the least skilled people who enjoyed the greatest benefits. The Homer Simpson of the world, long sidelined by technology, may finally find an invention on their side.

I am still unnerved by the damage that new generative AI systems could do to our already bruised information ecosystem and the upheaval they could cause in the world of knowledge work. But I’m also encouraged by the glimmer of hope that they could – could – improve the working lives of some long-marginalized people.

Homer Simpson famously proposed a toast: “To alcohol! The cause and solution of all life’s problems. Homers everywhere may soon feel the same way about ChatGPT.

Tim Harford’s children’s book, “The Truth Detective” (Wren & Rook), is now available

Follow @FTMag on Twitter to find out our latest stories first