Let us know about free updates
As artificial intelligence becomes more capable than ever, is it safe for some kind of job? “I’m sure the safest job in the world is probably the gardener,” confessed recently Martin Wolf, FT’s chief economics commentator. That seemed right. There are a few things that the computer can’t do.
The next morning, FT unveiled the “AI-bred Garden” and described intelligently automated drip irrigation, pest detectors, laser scarecrow systems, and solar-powered weeding robots. oof.
It is not entirely clear how much laser scarecrows and robotic weeds actually threaten the work of human gardeners, but prospects remind us that there is a distinction between work and tasks. Most jobs are bundles of interconnected tasks.
Gardeners need to do everything from cutting and weeding to diagnosing pest invasion, designing outdoor spaces, or communicating with the most difficult clients. Various AI systems can be useful for most of these tasks, but it does not mean that the gardener’s work will disappear, but rather that it will change shape.
The question is how do we change the shape of what we do with each new AI application? And do we like the reworking work available to us on the other side of this transformation?
Generation AI may be new, but these questions are not new. They date back to the Ludite protests in the early 1800s. Highly skilled textile workers saw the machines performing the most difficult parts of the job and were able to be replaced by low-wage workers with far less expertise.
And what’s the answer to those long-standing questions? They rely on both technology and work. There are lessons drawn from two contrasting precedents. These are digital spreadsheets and warehouse guidance earpieces such as “Jennifer Unit.” Digital spreadsheets that came to the market in 1979 immediately performed previous work by accountants, but accounting experts modeled a variety of scenarios and risks and moved towards more strategic and creative issues. Who doesn’t want a creative accountant?
The Jennifer Unit is a headset that guides a warehouse picker running around grabbing items from the shelf, tracking the final move and whispering your ears as you guide your next move. The unit removes the final trace of cognitive load from the physically demanding task that was already in the mind. This is in stark contrast to a digital spreadsheet that has removed the most boring parts of a variety of highly skilled jobs. Lesson: AI can make boring jobs even more boring and interesting jobs even more interesting.
New data and new perspectives on these questions come from MIT researchers David Autor and Neil Thompson. Autor and Thompson launch a new research paper entitled “Expertise” by raising questions. Do you expect accountants and inventory clerks to be affected by automation as well?
There are several established approaches to analyzing this question, all of which suggest that the answer is “yes”. At the time, both types of store clerks spent a lot of time performing everyday intellectual tasks, such as finding inconsistencies, editing tables of inventory and data, and doing simple arithmetic on a large scale. All of these tasks were like what a computer could do, and they took over as the computer was cheap enough. Given that the same task faces the same kind of automation, it appears logical that both jobs change in a similar way.
But that’s not what happened. In particular, Autor and Thompson say that the wages of accounting clerks have risen and the wages of inventory clerks have fallen.
This is because most jobs are not random collections of unrelated tasks. They are bundles of tasks that are most efficient by the same person for a variety of reasons. Removing some tasks from the bundle changes the rest of the job.
The inventory clerk lost a bit of work that required most education and training (arithmetic) and became like a shelf stacker. The accountants also lost arithmetic, but what remained required judgment, analysis and sophisticated problem resolution.
The same kind of tasks were automated, but the effect was to create jobs that required less training and expertise for the job that was invented, but the treasurer had to be more experts than before.
A natural concern for those hoping to work in five years is what AI does for the job. And while there is little certainty, the Autor and Thompson framework suggests clear questions. Does AI look like you’ll do the most highly skilled part of your job or the low-skilled ramp you couldn’t get rid of? The answer to that question may help you predict whether your job is making more enjoyable, or annoying, and whether your salary could rise, or whether professional jobs will be neglected like Luddites’ professional jobs.
For example, a generator AI system is excellent brainstorming. They create unexpected connections and generate many different ideas. When you’re running a role-playing game, that’s great. They accelerate their preparation and let’s go straight to the good stuff sitting around the table pretending to be a wizard with my friends.
Recommended
For those who offer an occasional oasis of creative brainstorming in the desert of work, the emergence of industrial brainstorming engines may be rather liberating.
First look at our latest stories – Follow FT Weekend Magazine on X and FT weekend on Instagram
