Bullsh*t Work in the AI Era

Are we automating away bullshit jobs – or about to create a whole lot more of them?

While Pascal and Jane were on an adventure to summit Mt. Kilimanjaro the past few weeks, I’ve been thinking about bullshit — specifically “bullshit jobs” in the sense that the late anthropologist David Graeber defined them. In 2013, Graeber wrote a quirky, provocative essay that later became a quirky, popular book, arguing that “huge swathes of people, in Europe and North America in particular, spend their entire working lives performing tasks they secretly believe do not really need to be performed.” Meaningless tasks. Busy work. Bullshit jobs. Graeber was exploring, in part, a question of technology and the different futures that new technologies could enable (or fail to enable), and his argument was also a lament that “despite our technological capacities, we are not all working 3 - 4 hour days” and have instead created an economy with “an apparently infinite demand for specialists in corporate law.” That and legions of managerial box-tickers and taskmasters.

You don’t have to buy all of the book’s claims (and its many generalizations) to recognize that Graeber had a point. Recently, I’ve been revisiting his theory of bullshit jobs in the context of conversations about the AI-enabled future of work, and I think we — today, collectively — might be at a pivotal moment in the history of bullshit jobs and busy work.

A few weeks ago, the Wall Street Journal ran a piece on the ”existential transformation” of McKinsey in a time when “artificial intelligence can increasingly do the work done by the firm’s highly paid consultants, often within minutes.” If I were a McKinsey guy (or someone who had paid a princely sum to bring them into my company), I don’t think I’d have been thrilled with the article, but as an external observer with an interest in the evolving value proposition of consulting and advisory work, I found it fascinating. Here was the WSJ asking not only how the biggest name in consulting manages to stay relevant in the Gen AI era but coming pretty close to asking a larger question about a lot of consulting work — and by extension, a lot of managerial/professional class jobs more generally. What is the actual value of the “value” they have been producing? And how might our perception of that value change when the work that supported it is automated? 

Consultants interviewed for the piece (at McKinsey and elsewhere) suggest that the key to maintaining their relevance — and revenues — will be a shift away from strategic advisory and deeper into implementation and projects with measurable, outcomes-based arrangements. Into the “trenches,” as one interviewee puts it. 

Has the Gen AI-enabled democratization of the slick PPT strategy deck arrived to liberate the consultants of McKinsey from bullshit work? Or at least the bullshit-ier aspects of their work? Maybe. 

Either way, I’m enthusiastic about the idea that new AI capabilities might push us to ask some overdue questions about the real value of certain services and tasks — and whether our time and energies might be more fruitfully directed elsewhere. And rather than just automating all of those services and tasks (and potentially producing more bullshit outputs than ever — a work slop-ocalypse, if you will), perhaps this AI moment, especially if it unfolds at a pace that allows us to be a little more intentional in what we automate and how, also offers us another chance to ask which work truly matters in the first place. 

@Jeffrey