Using generative AI to do the busy-work

Ray Fleming predicts:

"In ten years' time we will look back and wonder why we spent so much time doing paperwork, emails, applications and just plain dull processes. And how we managed to get any work done."

We're currently in an era marked by an incessant onslaught of administrative and repetitive tasks. Emails that require prompt responses, paperwork that seems to regenerate overnight, and processes that often feel like a maze designed to consume time. Now, let's catapult ourselves to a decade from now, to 2033. In that future, our perspective on work and productivity will likely be unrecognisable from today's norms. We will gaze back at our current workflows and be genuinely baffled. The question on everyone's lips will be, "Why did we spend so much time on routine work when we could have been breaking new ground in our respective fields?"

Automating the academic grind

In the realm of academia, the arduous process of grant application serves as a compelling example of the inefficiencies that plague modern workflows. According to an article in Nature last week, a staggering 90 to 95 percent of grant applications are destined for rejection. Each application mandates an exhaustive list of requirements, from project plans to justifications, benchmarks to historical data, all wrapped up in a labyrinth of documentation. And researchers are already finding that they can use generative AI, and specifically ChatGPT, to step in and shoulder this burden. Imagine if we built a specific tool set to do this, instead of a generalised one. These advanced AI systems wouldn't merely automate the writing process; they'd understand the essence of the proposed research, correlate that with historical data and funding requirements, and generate a grant application that significantly aligns with the criteria for success.

Baumol's Cost Disease: The unyielding inefficiency

In certain industries, there's a phenomenon that renders efforts at streamlining ineffective. This is known as Baumol’s Cost Disease, a theory that asserts that in some sectors, efficiency gains are hard or even impossible to achieve simply by scaling resources, and costs continue to escalate in line with wages in other industries. It was first diagnosed by Baumol and Bowen in their research in 1966, into performing arts. Classical music serves as an illustrative example. Doubling the number of musicians in an orchestra, or paying them more, doesn’t halve the time it takes to perform Beethoven’s Fifth Symphony; the symphony requires a set amount of time to deliver its full impact. Similarly, there are factors in the healthcare, education and legal sectors which create the same effect.

An important implication of Baumol effect is that it should be expected that, in a world with technological progress, the costs of manufactured goods will tend to fall (as productivity in manufacturing continually increases) while the costs of labor-intensive services like education, legal services, and health care (where productivity growth is persistently slow) will tend to rise

When generative AI changes the rules of efficiency

Against this backdrop, generative AI emerges as a remarkably potent solution. In the context of education, an industry where plenty has been written about the impact of Baumol's Cost Disease, the administrative tasks are often a time-consuming detour from the main road of imparting education. Just ponder the hours teachers spend on tasks such as marking, inputting grades, communicating with parents, and crafting lesson plans that match a rigid schema. Generative AI can be calibrated to execute these jobs seamlessly, leaving educators with more time for direct pedagogical engagement. The result is a profound shift in ‘time-on-task,’ wherein educators can focus on delivering teaching, potentially offering an avenue to counter Baumol’s Cost Disease (and making the role more fulfilling, as few teachers I know signed up to be spreadsheet wizards).

Generated with DALL-E 3 (and it took a while to convince it that robot teachers weren't the future)

The future is now

The implications of this coming tide of automation extend far beyond specific sectors or types of work. What we're looking at is a comprehensive recalibration of how we perceive and measure productivity. The real question is not whether this sea change is imminent—it clearly is. The question is how quickly we adapt and integrate these technological advancements into our daily lives. A strategic, well-thought-out approach to adopting AI for routine tasks can free us to focus on the aspects of our jobs that demand creativity, nuanced understanding, and human interaction.