Auto­ma­ti­on is a nice way to never inter­act with any­bo­dy. Buy­ing a drink from a ven­ding machi­ne, pay­ing by card at the petrol pump, or down­loa­ding some­thing from iTu­nes ins­te­ad of going to a record shop—they all have the bene­fit of making sure that you don’t have to deal with the vaga­ries and unpre­dic­ta­bi­li­ty of humans.


And the­se are just the out­liers. Rese­arch from the Nomu­ra Insti­tu­te, published in col­la­bo­ra­ti­on with the Uni­ver­si­ty of Oxford, posits that near­ly half of jobs in Japan will be auto­ma­ted in the next ten-to-twen­ty years. That num­ber makes for a sca­ry head­line.


But it’s also mis­lea­ding and the num­bers are open to inter­pre­ta­ti­on.


The­re are via­ducts in that logic, not just gaps. For star­ters, we don’t know how that 49 per cent is calculated—is it a pro­por­ti­on of pro­fes­si­ons or of the work­force? Tho­se are two very dif­fe­rent things. And the rese­arch does not take into account how an eco­no­my chan­ges so the peop­le who would have done tho­se soon-to-be-auto­ma­ted jobs may be freed to do some­thing more chal­len­ging or crea­ti­ve, crea­ting new pro­fes­si­ons.


And, as the South Chi­na Morning Post noted, “The stu­dy found tho­se jobs tend not to requi­re spe­cial know­ledge or skills.” So what will hap­pen is not that 49 per cent of jobs—that deba­ta­ble num­ber again—will be lost, but that auto­ma­ti­on will cut a swa­the through seg­ments of the unskil­led work­force.


So whe­re will this lea­ve jour­na­lism, a pro­fes­si­on that seems to bear the brunt of every tech­no­lo­gi­cal chan­ge? As a com­pa­ny spe­cia­li­sing in auto­ma­ted text, we have a vested inte­rest in how this will play out. But our posi­ti­on is this—a robot will beco­me a valu­able aide to jour­na­lists, not replace them.


One of our lan­guage archi­tects here at Ret­res­co was a tra­de hack in Lon­don in a pre­vious life. A lar­ge part of their job at the begin­ning of each year was to comb the annu­al state­ments of the major insu­rers, finan­ci­al insti­tu­ti­ons, and rep­re­sen­ta­ti­ve bodies to ana­ly­se and com­pa­re figu­res. It was a slog, but the­re were occa­sio­nal­ly some pearls. One year, they found that sales of a par­ti­cu­lar insuran­ce pro­duct had jum­ped by near­ly 460 per cent—a sto­ry that slip­ped lar­ge­ly under the radar.


So think of all the hours they spent sifting through that data. Our sys­tems can do that in the blink of an eye and find out whe­re tho­se bumps in the data exist, pin­poin­ting the way to major sto­ries that would other­wi­se go unre­por­ted. The­re is a say­ing that jour­na­lism is not just get­ting right the facts, but making the sure mea­ning of tho­se facts is cor­rect. That’s what our sys­tems do. We get the facts right so they can be inter­pre­ted cor­rect­ly.


Even the rese­ar­chers admit as much. Accord­ing to Arti­fi­cal Intel­li­gence Online, “[…] the new rese­arch shows that robots could per­form jobs which invol­ve more repe­ti­ti­ve tasks, and at the same time let humans pur­sue jobs that requi­re ana­ly­sing abs­tract thoughts, crea­ti­vi­ty, empa­thy, nego­tia­ti­on, and com­mu­ni­ca­ti­on – some­thing that robots would have a hard time doing.”


Being able to pro­du­ce auto­ma­ted reports on the mun­da­ne aspects of a reporter’s role means that that repor­ter can be freed to con­cen­tra­te on lon­ger-term, more-inves­ti­ga­ti­ve efforts such as the orig­ins of tho­se data bumps? And isn’t that what we real­ly need jour­na­lists to do?


Photo by Alex Knight on Unsplash