Policy experts and pundits are very excited about the idea that artificial intelligence (AI) will revolutionise the workplace. Perhaps it will, just not in the way they suppose.
For instead of making people redundant, ChatGPT may soon be empowering the most redundant people that your organisation employs.
A cartoon by Tom Fishburne captures this expectation. It shows two employees, each in their own office cubicle. “AI turns this single bullet point into a long email I can pretend I wrote,” says the first. Over in another part of the firm, a second employee enthuses: “AI makes a single bullet point out of this long email I pretend I read”.
If modern office work is really acting, then some employees really excel at looking busy. They are, as Bing Crosby once sang, “busy doing nothing … finding lots of things not to do”.
The core of the joke is that the latest AI chatbots will allow non-productive staff to look busier than ever: the AI is a sexy new prop for their performance. In turn, they will be able to devote more of their day to their personal activities at their employer’s expense, while giving them a ticket to the hottest technology show in town, one that will look great on their LinkedIn profile.
There’s much more idle and non-productive time at work than anyone wants to admit. But Fishburne’s cartoon also highlights another awkward truth. In the past, technology replaced human labour. Instead, information technology today creates and emboldens our workplace bureaucrats.
It’s been over three decades since the Nobel Prize-winning economist Robert Solow lamented: “You can see the computer age everywhere but in the productivity statistics.”
Since then, many major trends have swept through the labour market, each one promising to prove him wrong.
Thanks to personal computers, managers now type their own communications, rather than a temp.
Customer service is handled in cavernous offshore warehouses, or put back on the customer in the form of a self-service smartphone app.
Digital transformation was supposed to encompass all this, and also strip out administrative overhead from supply chains. Another great hype, “machine to machine” (M2M) communications, was also supposed to remove expensive human oversight and deliver us productivity savings.
So isn’t it time we cashed the cheque for some of these promises, and saw the results in our GDP and productivity figures? Alas, digital technology doesn’t seem to be delivering. Since the financial crash of 2008, there has been precious little labour productivity growth in the UK.
“We have a collective historical memory that technological progress brings a big and predictable stream of revenue growth across most of the economy. When it comes to the web, those assumptions are turning out to be wrong or misleading,” wrote Tyler Cowen a decade ago, in a pamphlet called The Great Stagnation.
Perhaps the savings made by deploying digital technology have been real, but simply swallowed up by non-productive work popping up somewhere else in the firm.
“It’s amazing that we’re as productive as we are, and half of what we do is compliance,” says Christopher Snowden of the Institute of Economic Affairs.
That’s a good place to start looking. Historically, an organisation had a department called “Personnel”, which did little more than collect references and process the payroll.
But in recent years, it has expanded its remit enormously, concerning itself with the employees health and “well-being”, which allowed it to intrude into how the employees behave.
The empire of Human Resources is very extensive indeed. HR will not only have crafted the employee handbook, but perhaps the mission and values statements too: things we never knew we needed until recently. The UK now employs over 400,000 people in human resources roles, almost as many people as work in agriculture. It provides a sinecure for the recent explosion of graduates, and a kind of secular manners and behaviour police.
Another culprit for the failure of technology to improve productivity may be, as Cowen suggests, the technology industry itself.
The venture capitalist Keith Rabois sparked a lively debate recently discussing the layoffs at large technology firms. “There’s nothing for these people to do … It’s all fake work,” he lamented. Over at Twitter, Elon Musk has dispensed with two thirds of its staff and contractors and the service remains much the same as it was before.
Into this delicate and complex sociological picture comes ChatGPT, the parrot that has swallowed the internet and can burp it back up again quite brilliantly. ChatGPT replicates the administrative formalities that consume a bureaucrat’s day.
While many of us would see this as an opportunity to identify and remove those redundant processes and posts, a canny idler spies the chance to entrench them. No productivity savings will be found if they succeed.
As Amazon’s CEO Andy Jassy pointed out last week, all this marvellous AI magic doesn’t come cheap. Today’s large language models cost billions of dollars to train. Silicon Valley will need to recoup that somehow. So we may well wake up to a day where our enterprises are no more productive than they were before, but must pay an additional service charge to Silicon Valley simply to keep their AI-toting bureaucrats looking busy.
Revolutions have started over less.
Add Comment