Slow robot’s

The march of the ma­chines has started, but it will take them some time to ar­rive

The Daily Telegraph - Business - - Front Page -

Ro­bots in space. In­tel­li­gent ma­chines in the freez­ing deeps of the North Sea. Au­toma­tons within the most toxic cham­bers of nu­clear power plants, or plung­ing into the harsh­est mines. The fu­ture for ar­ti­fi­cial in­tel­li­gence is ex­otic, alien and pow­er­ful.

At least, that is the vi­sion bought into by the Busi­ness De­part­ment last week when it ploughed an ex­tra £84m into a se­ries of new projects at a se­ries of uni­ver­si­ties and cor­po­rate re­search and de­vel­op­ment labs.

Cut­ting edge re­search into the ar­ti­fi­cial in­tel­li­gence, robotics, 3D print­ing and other barely-out-of-sci-fi fields holds plenty of prom­ise but has yet to com­pletely up­end the world in which we live.

For the past decade busi­ness in­vest­ment ap­pears to have been slug­gish. Pro­duc­tiv­ity growth has slowed and real wages are un­der the cosh, which does not fit with the idea that we are in an age of in­cred­i­ble tech­nolo­gies. Vi­sions of the fu­ture can be in­spir­ing or daunt­ing, de­pend­ing on your job right now.

Air­bus last week claimed it can have driver­less fly­ing cars on the mar­ket in five years’ time – a fan­tas­ti­cal vi­sion and the stuff of movies so far.

In per­haps a more grounded fore­cast, John Cryan, the chief ex­ec­u­tive of Deutsche Bank, said half of the bank’s al­most 100,000 jobs could be re­placed by ro­bots in the next 20 years.

So far, the prac­ti­cal gains have been more mod­est, but are im­por­tant in spe­cific ar­eas.

Daniel He­garty founded and runs Habito, an on­line mort­gage bro­ker which uses ar­ti­fi­cial in­tel­li­gence as part of its sys­tem to process ap­pli­ca­tions and match bor­row­ers with loans.

“Hu­mans are re­ally, re­ally bad at do­ing arith­metic and fill­ing out forms and hold­ing 80 dif­fer­ent credit poli­cies in their heads and match­ing them to the right cus­tomer,” he says.

There is still a hu­man el­e­ment – he found cus­tomers are un­will­ing to con­duct such an im­por­tant trans­ac­tion through a clever ma­chine alone, and value the chance to talk to a liv­ing, breath­ing ad­viser. “We use ma­chines for sort­ing and hu­mans for talk­ing,” he says. Nonethe­less, He­garty be­lieves he is four or five times more ef­fi­cient than a tra­di­tional bro­ker, with the ca­pac­ity to grow much fur­ther us­ing the tech­nol­ogy he al­ready has.

How­ever, he is aware these are early days in what is some­times called, rather grandly, the fourth in­dus­trial rev­o­lu­tion.

“We are su­per early on. We have es­sen­tially mech­a­nised a man­ual process. That is in­ter­est­ing and it pro­vides a tan­gi­ble con­sumer ben­e­fit. But the real trans­for­ma­tion in fi­nan­cial ser­vices is still some way off,” he says.

That be­lief is based on his con­cern that right now, peo­ple must shop around on a reg­u­lar ba­sis for new loans or sav­ings ac­counts with im­per­fect in­for­ma­tion, leav­ing them re­liant on the mar­ket­ing power of big fi­nance firms, rather than find­ing the best prod­uct out there.

“When ma­chines pick the prod­ucts it is not about the mar­ket­ing bud­get, it is about the best prod­uct,” he says, hop­ing it will force firms to com­pete on value. “That will be a trans­for­ma­tional point.”

It is not only ser­vices which could be rev­o­lu­tionised by new tech­nol­ogy. Man­u­fac­tur­ing, too, faces a wave of much-hyped new tech which is yet to fully hit home.

Econ­o­mist Raoul Leer­ing at ING has stud­ied the po­ten­tial of 3D print­ing and sees sev­eral fac­tors hold­ing it back so far.

“Cur­rent 3D print­ers are re­ally slow, so they are not com­pet­i­tive with re­gard to tra­di­tional cap­i­tal goods,” he says. “For the time be­ing, they are only com­pet­i­tive when mak­ing cus­tomised prod­ucts in a com­plex shape,” he says, cit­ing med­i­cal parts printed to match an in­di­vid­ual’s body. “But for mass pro­duc­tion of tra­di­tional prod­ucts, 3D print­ers are too slow. We are wait­ing for a new gen­er­a­tion to speed up the tempo.”

Ad­vances are quick – the lat­est mod­els are as much as 1000-times faster than their pre­de­ces­sors – but speed, price and re­li­a­bil­ity all have to com­bine to make them a worth­while pur­chase.

His ob­ser­va­tions in that mar­ket can be ap­plied to all rad­i­cal new tech.

Firstly, com­pa­nies are not go­ing to aban­don any ex­pen­sive rel­a­tively mod­ern equip­ment im­me­di­ately just be­cause the lat­est tech has marginally out­classed it.

Sec­ond, en­gi­neers with decades of ex­pe­ri­ence are not al­ways keen to aban­don all of the sys­tems with which they have ex­per­tise, switch­ing to a revo­lu­tion­ary new sys­tem overnight.

And third, he sees a net­work ef­fect hold­ing in­vest­ment back.

To use the ex­am­ple of the in­ter­net, the tech­nol­ogy took time to grow be­cause the first per­son with it could com­mu­ni­cate with no­body.

Go­ing back fur­ther in time, a lone tele­phone is use­less with no­body to call. It was only as more peo­ple gained ac­cess that it be­came worth any­body’s while to con­nect to the in­ter­net. And as more peo­ple did, ac­cess it­self be­came more valu­able and, in time in­dis­pens­able.

This slow adop­tion has raised com­par­isons with pre­vi­ous eras of slow pro­duc­tiv­ity and wage growth in Bri­tish his­tory.

“There has been a pe­riod, a 10-year pe­riod in the UK, when you hadn’t had any pro­duc­tiv­ity growth, but you have to go back to some time in the 1860s or 1870s to see this,” says Ben Broad­bent, deputy Gov­er­nor at the Bank of Eng­land.

“When peo­ple talk about that pe­riod, they talk about the pause be­tween two big tech­nolo­gies. We were mov­ing away from steam to elec­tric­ity.”

Ian Stewart, Deloitte’s chief econ­o­mist, says the adop­tion of elec­tric­ity into the work­place took far longer still than just that decade.

“Pro­duc­tiv­ity does not pro­ceed at a sta­ble rate over time. Af­ter the First World War par­tic­u­larly there was a mass de­ploy­ment of elec­tri­fi­ca­tion in fac­to­ries, so there was an ex­traor­di­nar­ily long lag from the in­ven­tions [in the 19th cen­tury] to their ap­pli­ca­tion in fac­to­ries,” he says. “The whole pro­duc­tion process had to be re­designed, build­ings had to be re­designed. And as that hap­pened there was a surge in pro­duc­tiv­ity.”

“It may be the case that we haven’t re­ally fully ex­ploited cur­rent tech­nolo­gies, let alone the next gen­er­a­tion.”

That is echoed by the CBI in a new re­port which notes that the UK has a small group of highly pro­duc­tive, cut­ting-edge firms, but a long tail of un­der­achiev­ers.

Al­most 70pc of em­ploy­ees in the UK work in com­pa­nies with be­low­me­dian pro­duc­tiv­ity, it notes, as too many firms have failed to adopt even well-es­tab­lished tech­nolo­gies.

While the slow pace of the AI rev­o­lu­tion and the rise of the ro­bots may be a drag for sci-fi fans, at least it means the most dire pre­dic­tions of mass re­dun­dan­cies and end­less poverty for swathes of the pop­u­la­tion are un­likely to come true as work­ers have time to adapt.

“It doesn’t re­place jobs, it frees work­ers up to do more high-value ac­tiv­ity, say in in­no­va­tion,” says Lee Ho­p­ley, chief econ­o­mist at man­u­fac­tur­ing in­dus­try group EEF.

The process is a slow one in any case, be­cause “this isn’t a big bang event where sud­denly these new tech­nolo­gies are avail­able off the shelf to ap­ply in your busi­ness,” she says.

Mar­ion Amiot at Ox­ford Eco­nom­ics says: “I don’t think the net ef­fect will be to de­stroy jobs.”

There is a clear de­sire to speed up in­vest­ment and the adop­tion of pro­duc­tiv­ity en­hanc­ing tech­nol­ogy. The Chan­cel­lor, Philip Ham­mond, is un­der pres­sure to find ways to raise in­vest­ment and boost the econ­omy in both the short and long-term.

But al­though there is no short­age of pro­pos­als sent his way, it is dif­fi­cult to find any per­fect an­swers. Broad­bent warns that the his­tor­i­cal record is not help­ful.

“It is not ev­i­dent that pol­i­cy­mak­ers can flick a switch and change this.

“If you ask most econ­o­mists they’d talk about sen­si­ble tax regimes, open­ness to the rest of the world, healthy pub­lic sec­tor in­vest­ment, good ed­u­ca­tion,” he says.

Deloitte’s Stewart also notes a dif­fer­ent les­son from his­tory: ex­pec­ta­tions can be hope­lessly wrong. He sees pro­duc­tiv­ity growth as a process which comes in waves, point­ing to the ter­ri­ble crunch in the Seven­ties, strong im­prove­ments as the econ­omy was re­vi­talised by the end of the Eight­ies and early Nineties, over-op­ti­mism with the dot­com bub­ble and now an­other flat pe­riod.

If this assess­ment is cor­rect, it may be that cur­rent pes­simism about the long-term fu­ture is se­verely over­done.

At least we can com­fort our­selves with that thought while we wait for space ro­bots to be­come a reg­u­lar fea­ture of our lives.

A 3D Prin­ter makes a plas­tic bowl, above; Air­bus be­lieves it will pro­duce driver­less fly­ing cars, be­low, in five years’ time, but the so-called fourth in­dus­trial rev­o­lu­tion will take time

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.