“You’ll never be able to replace me with a robot – they aren’t capable of the emotional connection that’s vital to my profession.” That’s the theory when it comes to the supplanting of certain positions with ever more sophisticated robots, but with the advent of a robot priest designed to spark debate about the future of the profession, it’s time to take a look at whether the same could happen to my own profession.
The fear of being replaced in your job by a more efficient successor is one of the oldest and most omnipresent societal fears. The fear of being replaced in it by a machine has existed since even before Fritz Lang made it a central component of Metropolis back in 1927.
It hasn’t been a fear confined to science fiction for centuries, with British workers as long ago as 1811 smashing the machines that they feared – correctly – would take their jobs. Nor, obviously, has that fear been assuaged by technological developments in the past few decades. If anything those fears are stoked more fiercely every year.
Recently attempts to automate aspects of the journalism process have become stories in their own right. Last year the AP began publishing automated stories in response to financial releases, through a process The Verge described as feeling like “a pretty standard, if a tad dry, AP news item. The obvious tell doesn’t come until the end of an article: ‘This story was generated by Automated Insights’.”
Since then there have been other examinations of what developments in artificial intelligence and machine learning could have on the process of journalism, beyond the minutiae of speeding up transcription and sourcing more relevant stories. A story published on The Guardian asked could AI ever win a Pulitzer, and our emeritus editor Damian Radcliffe asked “should we consider robo-journalists friends or foe?” back in July.
Speaking at FT Innovate Tunde Olanrewaju, senior partner at McKinsey & Company, examined “[as] the Daily Mail put it, ‘will a robot steal your job’?”
He cited the BBC’s ‘Will a robot take my job‘ calculator as a starting point. (‘Journalist, newspaper or periodical editor’, for the record, stands an 8 percent risk of being replaced within the next two decades, which was reassuring – though you have to consider the source).
He then described some examples of robots replacing jobs that have already happened:
“The first one is a chatbot that helps you deal with some very basic legal issues. Moley is still in production… a robot sous-chef, with a repertiore of over 2000 recipes. They have simulated how fingers work… recorded a chef doing what a chef does. You press play, and the robot basically replicates that. The next one is Sedaysys… this is live in the US. As you can expect the anaesthetist body in the US has seriously rebelled, Johnson & Johnson have taken it off the market.”
— Maria (@mariarmestre) November 2, 2016
Notably, the BBC’s tool gives anaesthetist (under the blanket term of ‘medical practictioner’) a 2 percent chance of being replaced by a robot within two decades. As Olanrewaju explains, it’s not necessarily a matter of robots being unable to perform the tasks – there are other factors in play. He highlighted three considerations in particular that McKinsey & Company examined to determine whether robots will see significant uptake within an industry:
- Who is disrupted
- Are they powerful enough to disrupt you in turn (through unions or in an industry-wide rebellion)
- Is it obvious to a consumer that the work is being done by a robot
Let’s look at those three factors with regards to journalists, many of whom must be especially worried about losing their jobs given rampant cost-cutting across the industry.
Who is disrupted?
So far the vast majority of the messaging around automated journalism has highlighted the benefits to journalists, rather than any disruption. Bloomberg’s editor-in-chief John Micklethwait said, in an internal memo earlier this year: “Done properly, automated journalism has the potential to make all our jobs more interesting.“
The idea is that the bare bones, rote level of reporting – largely that which is numbers-based – can be left to robots, freeing up reporters and writers to do the more interesting, analysis-led part of the job. So yes, it will be the journalists that are disrupted but, as assistant business editor at the AP Philippa Patterson explains, that’s not necessarily a bad thing.
“One of the things we really wanted reporters to be able to do was when earnings came out to not have to focus on the initial numbers. That’s the goal, to write smarter pieces and more interesting stories.”
Can journalists disrupt robots in turn?
This is uncertain. Journalists are already habituated to some extent to robots and automated processes aiding them with their jobs, so it’s highly unlikely there would ever be an industry-wide rebellion against the introduction of endeavours like that at the AP, particularly when there are more fun distractions on which to focus.
Instead it’s more likely we’ll see continued protests at the corollary effects of the economic climes that would encourage media companies to adopt automated journalists in the first place, such as job cuts and – unfortunately given what we’ll talk about later – pay and conditions.
Is it obvious the work is being done by a robot?
As The Verge suggests, you might not know that a piece of automated financial news was generated by a robot at first blush or that a picture caption was automatically generated, but given the current limitations of machine learning and AI it seems certain to be a while before a bot can pass the Turing test and write a listicle on BuzzFeed, much less an in-depth piece of geopolitical analysis.
So it all suggests that journalists won’t be outright replaced by robots in the way that many manufacturing jobs might be. Instead, as Olanrewaju explains:
“A robot will not steal your job – a robot will share your job.”
So, good news for journalists? Not necessarily. Olanrewaju and researchers at McKinsey & Company took a look at what automation of jobs has done for wages. Dividing the industries they surveyed into blue-collar, grey-collar and white-collar jobs (into which journalism arguably fits), they found that the overall number of blue-collar jobs has decreased as a result of automation, while the number of white- and grey-collar jobs has increased.
Unfortunately, that sharing of a role with automated processes has lessened the perceived value of workers at the low-end of the pay scale for white-collar jobs, to the point that they are only paid a little more than the low-end of blue-collar workers, a stark contrast to the situation at the turn of the century.
“If you index the bottom of white collar pay to 100, how much is the disparity? That gap has widened by 7x, the bottom end of white collar is approaching blue collar. Yes there are more jobs, but those jobs are paying a lot less. This is probably the effect of the automation.”
So while journalists are highly unlikely to be completely replaced by automated processes any time in the near future, the automated processes that make their jobs easier are likely to contribute to lower wages for the journalists at the start of their careers. As John Thornhill, innovation editor for The Financial Times noted: “We are not yet obsolete wetware” – but we might be devalued instead.